Introduction: For the treatment of single-sided deafness (SSD), common treatment choices include a contralateral routing of signals (CROS) hearing aid, a bone conduction device (BCD), and a cochlear implant (CI). The primary aim of this study was to compare speech understanding in noise and binaural benefits in adults with postlingual SSD between preoperative unaided baseline, preoperative CROS and BCD trial devices, and CI, following recommendations from a consensus protocol. In addition, we investigated the effect of masker type on speech understanding. Methods: This was a prospective study with twelve participants. Binaural effects of head shadow, squelch, summation, and spatial release from masking were assessed by measuring speech reception thresholds (SRTs) in five different spatial target-masker configurations using two different maskers: two-talker babble (TTB), and speech-shaped noise (SSN). Preoperatively, participants were assessed unaided and with CROS and BCD trial devices. After cochlear implantation, participants were assessed at 1, 3, and 6 months post-activation. Results: For TTB, significant improvements in SRT with a CI relative to preoperatively unaided were found in all spatial configurations. With CI at 6 months, median benefits were 7.8 dB in SSSDNAH and 5.1 dB in S0NAH (head shadow), 3.4 dB in S0N0 (summation), and 4.6 dB in S0NSSD and 5.1 dB in SAHNSSD (squelch). CROS yielded a significant head shadow benefit of 2.4 dB in SSSDNAH and a significant deterioration in squelch of 2.5 dB in S0NSSD and SAHNSSD, but no summation effect. With BCD, there was a significant summation benefit of 1.5 dB, but no head shadow nor squelch effect. For SSN, significant improvements in SRT with CI compared to preoperatively unaided were found in three spatial configurations. Median benefits with CI at 6 months were: 8.5 dB in SSSDNAH and 4.6 dB in S0NAH (head shadow), 1.4 dB in S0N0 (summation), but no squelch. CROS showed a significant head shadow benefit of 1.7 dB in SSSDNAH, but no summation effect, and a significant deterioration in squelch of 2.9 dB in S0NSSD and 3.2 dB in SAHNSSD. With BCD, no binaural effect was obtained. Longitudinally, we found significant head shadow benefits with a CI in SSSDNAH in both maskers at all postoperative intervals and in S0NAH at 3 and 6 months post-activation. Conclusion: With a CI, a clear benefit for masked speech perception was observed for all binaural effects. Benefits with CROS and BCD were more limited. CROS usage was detrimental to the squelch effect.

The primary hearing problems encountered by people with single-sided deafness (SSD) or asymmetric hearing loss (AHL) are difficulties in understanding speech in challenging listening situations, particularly in noisy settings [1, 2], and in localizing sound sources [3, 4]. According to Van de Heyning and colleagues [5], SSD is defined as having a four-frequency (0.5, 1.0, 2.0, and 4.0 kHz) pure-tone threshold average (4PTA) of ≥70 dB HL in the poorer ear and ≤30 dB HL in the better ear, with an interaural threshold gap of ≥40 dB HL. AHL is defined as having a 4PTA of ≥70 dB HL in the poorer ear and between >30 dB and ≤55 dB HL in the better ear with an interaural threshold gap of ≥15 dB HL [5].

Cochlear implant (CI) provision is an effective clinical intervention in individuals with SSD or AHL. In such cases, a CI allows for the (re)habilitation of binaural hearing through electrical stimulation of the ear with severe-to-profound hearing loss [5‒7]. The CI use in people with SSD and AHL provides significantly better speech understanding in noise [7‒11] and sound localization [7, 9, 12, 13] than either nonintervention or intervention with a contralateral routing of signals (CROS) hearing aid or a bone conduction device (BCD) [7, 9, 11]. Moreover, the CI use has been shown to significantly reduce or eliminate tinnitus, a common issue with SSD [14‒18], and to improve users’ quality of life [6, 12, 19‒22].

However, the results of past studies are rarely comparable due to varying inclusion criteria and differences in testing methodologies and conditions. These include differences in intervals, setups, stimuli, as well as in measurement procedures for the assessment of speech understanding, sound localization, tinnitus, and quality of life [5, 23]. Another issue is that the number of study participants is often relatively small; therefore, drawing robust statistical conclusions can be challenging and results might not be directly representative of population outcomes.

To resolve this, an international committee of research groups proposed a unified testing framework that, if followed, would harmonize study design and minimum outcome measures across studies [5]. This would allow studies using this framework to be more directly comparable. This has the advantage of further enabling data from different centers to be pooled, thus allowing more robust conclusions to be drawn, and better overall decision-making regarding treatment.

The first study that followed the unified testing framework had a retrospective design and was published by the Comprehensive Hearing Center at the University of Würzburg [7]. The present study is, to the best of our knowledge, the first to implement this framework in a prospective design. The primary aim of this study was to evaluate the longitudinal speech understanding and binaural benefits of CI use in people with SSD and to compare these to the outcomes achieved with CROS and BCD trial devices. We evaluated masked speech perception using several spatial target-masker configurations in order to investigate different binaural effects, namely: head shadow, summation, squelch, and spatial release from masking (SRM). A secondary aim was to evaluate the effect of masker type on speech understanding. Two different maskers were applied: two-talker babble (TTB) and speech-shaped noise (SSN). Both maskers are commonly used in speech perception testing, but there is reason to believe that they differ in terms of the manner in which they mask target speech identification. This study allowed the direct comparison of these two maskers for speech understanding with different hearing devices.

This prospective study was approved by the Ethics Committee of the University of Freiburg, Germany (protocol number 230/16). All procedures were performed and conducted in accordance with the recommendations of this Committee and in accordance with the Declaration of Helsinki (2013). All participants gave their informed written consent prior to the beginning of the study.

Inclusion and Exclusion Criteria

To be included in the study, potential participants had to meet all of the following inclusion criteria: be an adult (18 years of age or older); be proficient in the German language; and have postlingual SSD or AHL, as defined by Van de Heyning et al. [5]. In the case of a borderline PTA in the poorer hearing ear, participants were included if they had a best-aided preoperative monosyllabic word score of less than 50%. Exclusion criteria consisted of having either a cognitive impairment or a cochlear malformation in the poorer hearing ear. Consecutive patients with SSD from clinical routine examinations were counseled on the treatment options of CROS, BCD, CI, and no treatment. Patients considering a CROS, BCD, or a MED-EL CI in case of a decision to undergo cochlear implantation were included in the study.

Study Design

Participants were assessed at three preoperative intervals and three postoperative intervals.

Preoperative investigation: At the first preoperative interval, participants were tested in the bilaterally unaided condition. For the subsequent preoperative CROS and BCD trials, participants were randomized into two groups. Participants in group A were provided and fitted with a loaner CROS hearing aid (Phonak CROS II-13, Bolero V90) for the first 3 weeks, followed by a second 3-week trial of a loaner BCD (Oticon Ponto Plus Power at headband). Participants in group B were provided and fitted with the same CROS and BCD, in the opposite order. At the end of each trial phase, i.e., the second and third preoperative interval, participants were tested in the aided condition.

Postoperative investigation: After the preoperative investigation, participants were offered the choice of treatment with either a CROS, BCD, or CI or no treatment. All participants who opted for a MED-EL CI were assessed at three postoperative intervals: 1, 3, and 6 months post-activation of the CI sound processor. At these intervals, participants were tested in the CI-aided condition.

Speech Understanding in Noise

At all intervals, speech-in-noise perception was quantified as the speech reception threshold (SRT) and assessed using the Oldenburg sentence test (OLSA) [24]. The OLSA is a matrix test with a German native male target speaker. SRTs were measured in two different maskers: (1) in a TTB consisting of meaningful German everyday sentences spoken by two male speakers who were not the target speaker, and (2) in the original OLSA masker OLnoise, a SSN matching the long-term spectrum of the target sentences. For SRT assessment, the masker was fixed at a level of 65 dB(A) while the level of the speech was adaptively modified in a stepwise procedure to converge at 50% correct word understanding.

For both the TTB and SSN maskers, SRTs were measured in five spatial configurations: speech presented from the deaf side at ±90° and the masker from the acoustic-hearing side at ∓90° (SSSDNAH); speech from the acoustic-hearing side and the masker from the deaf side (SAHNSSD); and three configurations with speech from the front (0°), where the masker was either presented from the acoustic-hearing side (S0NAH), from the front (S0N0), or from the deaf side (S0NSSD). Participants were seated in a sound-proof room at the center of a semicircular array of three loudspeakers, placed at head-level at 0°, 90°, and −90° azimuth, and a radial distance of 1 m from the center of the participant’s head. At each interval, prior to testing, four OLSA practice lists were presented: the first two in quiet, followed by one list in TTB and one list in SSN. To minimize potential bias, the order of the presentation conditions was randomized for each masker, and the two maskers were alternated.

For all treatment conditions including preoperatively unaided, CROS and BCD, and CI at 6 months, the following binaural effects were assessed for speech understanding in noise: head shadow, summation, squelch, and SRM. Head shadow, summation, and squelch were computed as differences in SRTs between the preoperatively unaided condition and aided conditions in SSSDNAH and S0NAH, in S0N0, and in S0NSSD and SAHNSSD, respectively. SRM was computed as differences in aided SRTs between S0N0 and S0NSSD and between S0N0 and S0NAH. Positive differences indicate benefit; negative differences indicate deterioration.

Statistical Analysis

Statistical analysis was performed with IBM SPSS Statistics 27 (IBM, Armonk, NY, USA). The level of significance was set to 0.05.

Friedman tests were used to assess the effect of the treatment condition on SRT in all maskers and spatial configurations. Post hoc Wilcoxon signed-rank tests with Bonferroni-Holm corrections for multiple comparisons were applied to compare SRTs between the treatment conditions. These tests included comparisons between SRTs for each device condition (CROS and BCD, and CI at 6 months) and the unaided condition, which correspond to the binaural effects of head shadow, summation, and squelch. To evaluate the binaural effect of SRM for significance for each device condition and both maskers, one-sample Wilcoxon signed-rank tests were used. Friedman tests with post hoc two-sample Wilcoxon signed-rank tests and Bonferroni-Holm corrections were applied to compare SRMs between device conditions. To evaluate speech understanding in both maskers and each spatial configuration over time, SRTs for the preoperatively unaided condition, and with CI at 1, 3, and 6 months, were compared with Friedman tests and post hoc Wilcoxon signed-rank tests with Bonferroni-Holm corrections.

Group differences in SRT between treatment and device conditions were expressed as median benefit or deterioration and were calculated as medians of the individual differences between conditions. Binaural effect sizes were expressed as median benefit or deterioration and were calculated as medians of the individual differences in SRT for the respective spatial configurations.

For the preoperatively unaided condition, CI at 6 months, and the benefits with CI at 6 months compared to preoperatively unaided, the effects of masker and spatial configuration on SRT, respectively, SRT benefits were assessed using separate two-way repeated-measures ANOVAs. Post hoc paired t tests with Bonferroni-Holm corrections for multiple comparisons were applied to compare masker conditions.

Participants

Thirteen subjects (4 female, 9 male) were included in the study. One participant chose to discontinue study participation during the preoperative trial phase and was excluded from data analysis. At inclusion, ten of the twelve participants were neither using a hearing aid (ipsilateral or contralateral) nor a (Bi)CROS device. Two participants (P6, P10) wore an ipsilateral hearing aid. All twelve participants, six in either group A and B, completed the preoperative trials, opted for treatment with a MED-EL CI, and reached the study endpoint. Eleven participants had severe-to-profound hearing loss and one (P5) had moderately severe hearing loss in the ipsilateral ear. Their median age at inclusion was 52.5 years (range: 44.0–72.8 years) and the median duration of unilateral hearing loss was 3.5 years (range: 0.6–12 years). Additional demographic information is presented in Table 1. Preoperative audiograms of both ears are shown in Figure 1.

Table 1.

Demographics of study participants

ParticipantAge at inclusion, years4PTA of nonimplanted ear, dB HLImplanted ear
4PTA, dB HLbest-aided monosyllabic word score @ 65 dB SPL, %sideetiologyduration of severe to profound hearing loss, yearsimplant electrode arrayaudio processor
P1 53.5 17.8 83.5 SSHL SYNCHRONY FLEX 28 SONNET 
P2 62.3 23.0 87.8 SSHL SYNCHRONY FLEX 24 RONDO 2 
P3 72.8 26.3 77.8 SSHL 0.8 SYNCHRONY FLEX 28 SONNET 
P4 52.2 13.3 86.3 SSHL 0.6 SYNCHRONY FLEX 28 SONNET 
P5 64.8 13.3 65.8 SSHL 9* SYNCHRONY FLEX 28 SONNET 
P6 44.0 3.3 70.8 SSHL 0.8 SYNCHRONY FLEX 24 SONNET 
P7 46.0 16.0 72.8 Intracochlear schwannoma SYNCHRONY FLEX 24 SONNET 
P8 65.0 18.3 101.3 SSHL SYNCHRONY FLEX 28 RONDO 2 
P9 52.7 16.8 115.0 Vestibular schwannoma 12 SYNCHRONY FLEX 28 SONNET 
P10 46.8 20.3 94.0 20 SSHL SYNCHRONY FLEX 24 RONDO 2 
P11 48.0 18.3 113.0 Endolymphatic hydrops SYNCHRONY FLEX 28 RONDO 2 
P12 52.3 20.3 93.3 SSHL SYNCHRONY FLEX 28 SONNET 
ParticipantAge at inclusion, years4PTA of nonimplanted ear, dB HLImplanted ear
4PTA, dB HLbest-aided monosyllabic word score @ 65 dB SPL, %sideetiologyduration of severe to profound hearing loss, yearsimplant electrode arrayaudio processor
P1 53.5 17.8 83.5 SSHL SYNCHRONY FLEX 28 SONNET 
P2 62.3 23.0 87.8 SSHL SYNCHRONY FLEX 24 RONDO 2 
P3 72.8 26.3 77.8 SSHL 0.8 SYNCHRONY FLEX 28 SONNET 
P4 52.2 13.3 86.3 SSHL 0.6 SYNCHRONY FLEX 28 SONNET 
P5 64.8 13.3 65.8 SSHL 9* SYNCHRONY FLEX 28 SONNET 
P6 44.0 3.3 70.8 SSHL 0.8 SYNCHRONY FLEX 24 SONNET 
P7 46.0 16.0 72.8 Intracochlear schwannoma SYNCHRONY FLEX 24 SONNET 
P8 65.0 18.3 101.3 SSHL SYNCHRONY FLEX 28 RONDO 2 
P9 52.7 16.8 115.0 Vestibular schwannoma 12 SYNCHRONY FLEX 28 SONNET 
P10 46.8 20.3 94.0 20 SSHL SYNCHRONY FLEX 24 RONDO 2 
P11 48.0 18.3 113.0 Endolymphatic hydrops SYNCHRONY FLEX 28 RONDO 2 
P12 52.3 20.3 93.3 SSHL SYNCHRONY FLEX 28 SONNET 

For calculating the four-frequency pure-tone average (4PTA), unmeasurable thresholds were replaced by the frequency-specific maximum output limits of the audiometer plus 5 dB HL.

SSHL, sudden sensorineural hearing loss; R, right; L, left.

*Participant with moderately severe hearing loss.

Fig. 1.

Preoperative air-conduction hearing thresholds of the acoustic-hearing ear (left panel) and the deaf ear (right panel). Unmeasurable thresholds were replaced by the frequency-specific maximum output limits of the audiometer plus 5 dB HL.

Fig. 1.

Preoperative air-conduction hearing thresholds of the acoustic-hearing ear (left panel) and the deaf ear (right panel). Unmeasurable thresholds were replaced by the frequency-specific maximum output limits of the audiometer plus 5 dB HL.

Close modal

Speech Understanding in Noise

Binaural Effects with CROS/BCD Trial Devices and CI

Figure 2 presents SRTs in noise across spatial configurations and maskers for all listening conditions including preoperatively unaided, CROS, BCD, and CI at 1, 3, and 6 months. For both maskers, the effect of treatment condition on SRT was analyzed independently for the five spatial configurations. Differences in SRT between device conditions and the preoperatively unaided condition for each of the spatial configurations represent binaural effects of head shadow (SSSDNAH, S0NAH), summation (S0N0), and squelch (S0NSSD, SAHNSSD). Thus, SRT differences between device conditions (CROS and BCD trial devices, and CI at 6 months) correspond to differences in binaural effects between devices. Results from the respective Friedman tests and post hoc multiple comparisons are presented in Table 2.

Fig. 2.

Box and whisker plots of speech reception thresholds (SRTs) in each masker – two-talker babble (TTB, top panels) and speech-shaped noise (SSN, bottom panels) – and spatial configuration – SSSDNAH, S0NAH, S0N0, S0NSSD, SAHNSSD – tested. Each panel presents results for all listening conditions: preoperatively unaided, CROS and BCD trials; and CI at 1, 3, and 6 months. The boxes show interquartile range (IQR), the whiskers show the data range excluding outliers, where outliers are defined as values further out than 1.5 times IQR from the outer box edges. Solid lines within the boxes represent the median and diamonds represent the mean. Brackets with asterisks above the box plots represent significant differences in performance between treatment conditions and brackets below significant changes over time. *p < 0.05.

Fig. 2.

Box and whisker plots of speech reception thresholds (SRTs) in each masker – two-talker babble (TTB, top panels) and speech-shaped noise (SSN, bottom panels) – and spatial configuration – SSSDNAH, S0NAH, S0N0, S0NSSD, SAHNSSD – tested. Each panel presents results for all listening conditions: preoperatively unaided, CROS and BCD trials; and CI at 1, 3, and 6 months. The boxes show interquartile range (IQR), the whiskers show the data range excluding outliers, where outliers are defined as values further out than 1.5 times IQR from the outer box edges. Solid lines within the boxes represent the median and diamonds represent the mean. Brackets with asterisks above the box plots represent significant differences in performance between treatment conditions and brackets below significant changes over time. *p < 0.05.

Close modal
Table 2.

Speech understanding in two-talker babble (TTB) and speech-shaped noise (SSN) for each of the four treatment conditions (preoperatively unaided, CROS and BCD trials, CI at 6 months) and five spatial configurations (SSSDNAH, S0NAH, S0N0, S0NSSD, SAHNSSD) tested

Median SRT (dB SNR)Friedman testMedian and range of SRT difference (dB)
preoperatively unaidedCROSBCDCI 6mCROS versus preoperatively unaidedBCD versus preoperatively unaidedCI 6m versus preoperatively unaidedCI 6m versus CROSCI 6m versus BCDCROS versus BCD
TTB 
 Head shadow 
  SSSDNAH 0.2 −2.3 −1.0 −8.2 χ2(3) = 26.2 2.4 (−2.4–8.1) 1.4 (−2.3–8.9) 7.8 (2.4–12.0) 5.8 (0.2–10.2) 6.2 (−3.7–13.3) 1.6 (−6.0–4.2) 
p < 0.001 Z = −2.590 Z = −2.080 Z = −3.061 Z = −3.059 Z = −2.824 Z = −0.628 
 padj = 0.029 padj = 0.075 padj = 0.013 padj = 0.011 padj = 0.019 padj = 0.530 
  S0NAH −3.3 −2.6 −3.6 −8.0 χ2(3) = 18.5 0.5 (−4.7–3.8) 1.2 (−3.6–4.3) 5.1 (1.3–11.3) 4.9 (−1.6–12.5) 4.2 (−0.6–11.6) −0.6 (−6.1–2.8) 
p < 0.001 Z = −0.235 Z = −1.452 Z = −3.059 Z = −2.981 Z = −2.981 Z = −1.255 
 padj = 0.814 padj = 0.439 padj = 0.013 padj = 0.014 padj = 0.014 padj = 0.419 
 Summation 
  S0N0 −3.6 −4.3 −4.9 −7.3 χ2(3) = 22.1 1.6 (−3.1–3.6) 1.5 (−2.3–4.0) 3.4 (2.9–5.3) 1.9 (−0.1–6.6) 1.6 (−1.0–6.4) −0.4 (−4.6–3.0) 
p < 0.001 Z = −1.957 Z = −2.510 Z = −3.061 Z = −2.981 Z = −2.824 Z = −0.628 
 padj = 0.101 padj = 0.036 padj = 0.013 padj = 0.014 padj = 0.019 padj = 0.530 
 Squelch 
  S0NSSD −9.0 −7.0 −7.6 −12.6 χ2(3) = 22.8 −2.5 (−8.0–1.9) 0.2 (−5.8–4.4) 4.6 (−1.8–9.0) 6.7 (0.2–11.7) 4.1 (0.6–11.7) −2.0 (−5.8–3.7) 
p < 0.001 Z = −2.747 Z = −0.549 Z = −2.667 Z = −3.059 Z = −3.059 Z = −1.804 
 padj = 0.024 padj = 0.583 padj = 0.023 padj = 0.013 padj = 0.013 padj = 0.142 
  SAHNSSD −13.6 −10.1 −12.3 −17.7 χ2(3) = 13.3 −2.5 (−7.5–0.9) −0.8 (−4.5–2.6) 5.1 (−3.6–10.6) 8.6 (−4.5–11.4) 6.3 (−5.3–9.0) −2.5 (−5.8–4.3) 
p = 0.004 Z = −2.825 Z = −1.099 Z = −2.432 Z = −2.824 Z = −2.590 Z = −2.040 
 padj = 0.028 padj = 0.272 padj = 0.045 padj = 0.024 padj = 0.038 padj = 0.083 
SSN 
 Head shadow 
  SSSDNAH 1.9 −1.0 0.9 −8.2 χ2(3) = 24.3 1.7 (−1.0–5.5) 1.0 (−3.5–4.1) 8.5 (3.3–13.0) 6.6 (1.1–10.7) 8.1 (−0.8–12.8) 1.6 (−2.6–3.6) 
p < 0.001 Z = −2.510 Z = −1.334 Z = −3.059 Z = −3.064 Z = −2.981 Z = −2.040 
 padj = 0.036 padj = 0.182 padj = 0.011 padj = 0.013 padj = 0.011 padj = 0.083 
  S0NAH −1.1 −2.4 −2.5 −5.9 χ2(3) = 20.5 0.2 (−3.1–2.9) 1.1 (−3.4–3.3) 4.6 (1.0–7.5) 3.8 (1.2–8.4) 3.9 (−0.5–7.5) −0.3 (−2.5–3.8) 
p < 0.001 Z = −0.471 Z = −0.981 Z = −3.061 Z = −3.059 Z = −2.982 Z = −0.589 
 padj = 0.638 padj = 0.979 padj = 0.013 padj = 0.011 padj = 0.011 padj = 1 
 Summation 
  S0N0 −4.7 −4.6 −4.9 −5.8 χ2(3) = 19.6 0.3 (−1.4–1.5) 0.4 (−2.5–1.6) 1.4 (0.2–2.8) 1.4 (−0.2–1.7) 0.9 (0.0–3.3) −0.4 (−1.4–1.9) 
p < 0.001 Z = −0.981 Z = −1.335 Z = −3.062 Z = −2.983 Z = −2.809 Z = −0.589 
 padj = 0.653 padj = 0.546 padj = 0.013 padj = 0.014 padj = 0.020 padj = 0.556 
 Squelch 
  S0NSSD −9.6 −7.5 −8.8 −10.5 χ2(3) = 18.5 −2.9 (−6.9–2.3) −0.4 (−7.9–2.8) 2.2 (−3.6–4.2) 4.0 (1.9–6.7) 1.8 (−1.5–7.8) −1.5 (−5.4–3.3) 
p < 0.001 Z = −2.589 Z = −0.746 Z = −1.883 Z = −3.059 Z = −2.669 Z = −2.001 
 padj = 0.039 padj = 0.456 padj = 0.119 padj = 0.013 padj = 0.038 padj = 0.136 
  SAHNSSD −13.6 −10.8 −12.3 −14.6 χ2(3) = 24.9 −3.2 (−4.8–1.1) −1.4 (−5.8–2.0) 0.6 (−1.6–4.2) 3.4 (1.4–6.7) 1.4 (0.2–6.5) −1.9 (−3.4–2.2) 
p < 0.001 Z = −2.903 Z = −1.884 Z = −1.883 Z = −3.061 Z = −3.061 Z = −2.312 
 padj = 0.015 padj = 0.119 padj = 0.060 padj = 0.013 padj = 0.013 padj = 0.062 
Median SRT (dB SNR)Friedman testMedian and range of SRT difference (dB)
preoperatively unaidedCROSBCDCI 6mCROS versus preoperatively unaidedBCD versus preoperatively unaidedCI 6m versus preoperatively unaidedCI 6m versus CROSCI 6m versus BCDCROS versus BCD
TTB 
 Head shadow 
  SSSDNAH 0.2 −2.3 −1.0 −8.2 χ2(3) = 26.2 2.4 (−2.4–8.1) 1.4 (−2.3–8.9) 7.8 (2.4–12.0) 5.8 (0.2–10.2) 6.2 (−3.7–13.3) 1.6 (−6.0–4.2) 
p < 0.001 Z = −2.590 Z = −2.080 Z = −3.061 Z = −3.059 Z = −2.824 Z = −0.628 
 padj = 0.029 padj = 0.075 padj = 0.013 padj = 0.011 padj = 0.019 padj = 0.530 
  S0NAH −3.3 −2.6 −3.6 −8.0 χ2(3) = 18.5 0.5 (−4.7–3.8) 1.2 (−3.6–4.3) 5.1 (1.3–11.3) 4.9 (−1.6–12.5) 4.2 (−0.6–11.6) −0.6 (−6.1–2.8) 
p < 0.001 Z = −0.235 Z = −1.452 Z = −3.059 Z = −2.981 Z = −2.981 Z = −1.255 
 padj = 0.814 padj = 0.439 padj = 0.013 padj = 0.014 padj = 0.014 padj = 0.419 
 Summation 
  S0N0 −3.6 −4.3 −4.9 −7.3 χ2(3) = 22.1 1.6 (−3.1–3.6) 1.5 (−2.3–4.0) 3.4 (2.9–5.3) 1.9 (−0.1–6.6) 1.6 (−1.0–6.4) −0.4 (−4.6–3.0) 
p < 0.001 Z = −1.957 Z = −2.510 Z = −3.061 Z = −2.981 Z = −2.824 Z = −0.628 
 padj = 0.101 padj = 0.036 padj = 0.013 padj = 0.014 padj = 0.019 padj = 0.530 
 Squelch 
  S0NSSD −9.0 −7.0 −7.6 −12.6 χ2(3) = 22.8 −2.5 (−8.0–1.9) 0.2 (−5.8–4.4) 4.6 (−1.8–9.0) 6.7 (0.2–11.7) 4.1 (0.6–11.7) −2.0 (−5.8–3.7) 
p < 0.001 Z = −2.747 Z = −0.549 Z = −2.667 Z = −3.059 Z = −3.059 Z = −1.804 
 padj = 0.024 padj = 0.583 padj = 0.023 padj = 0.013 padj = 0.013 padj = 0.142 
  SAHNSSD −13.6 −10.1 −12.3 −17.7 χ2(3) = 13.3 −2.5 (−7.5–0.9) −0.8 (−4.5–2.6) 5.1 (−3.6–10.6) 8.6 (−4.5–11.4) 6.3 (−5.3–9.0) −2.5 (−5.8–4.3) 
p = 0.004 Z = −2.825 Z = −1.099 Z = −2.432 Z = −2.824 Z = −2.590 Z = −2.040 
 padj = 0.028 padj = 0.272 padj = 0.045 padj = 0.024 padj = 0.038 padj = 0.083 
SSN 
 Head shadow 
  SSSDNAH 1.9 −1.0 0.9 −8.2 χ2(3) = 24.3 1.7 (−1.0–5.5) 1.0 (−3.5–4.1) 8.5 (3.3–13.0) 6.6 (1.1–10.7) 8.1 (−0.8–12.8) 1.6 (−2.6–3.6) 
p < 0.001 Z = −2.510 Z = −1.334 Z = −3.059 Z = −3.064 Z = −2.981 Z = −2.040 
 padj = 0.036 padj = 0.182 padj = 0.011 padj = 0.013 padj = 0.011 padj = 0.083 
  S0NAH −1.1 −2.4 −2.5 −5.9 χ2(3) = 20.5 0.2 (−3.1–2.9) 1.1 (−3.4–3.3) 4.6 (1.0–7.5) 3.8 (1.2–8.4) 3.9 (−0.5–7.5) −0.3 (−2.5–3.8) 
p < 0.001 Z = −0.471 Z = −0.981 Z = −3.061 Z = −3.059 Z = −2.982 Z = −0.589 
 padj = 0.638 padj = 0.979 padj = 0.013 padj = 0.011 padj = 0.011 padj = 1 
 Summation 
  S0N0 −4.7 −4.6 −4.9 −5.8 χ2(3) = 19.6 0.3 (−1.4–1.5) 0.4 (−2.5–1.6) 1.4 (0.2–2.8) 1.4 (−0.2–1.7) 0.9 (0.0–3.3) −0.4 (−1.4–1.9) 
p < 0.001 Z = −0.981 Z = −1.335 Z = −3.062 Z = −2.983 Z = −2.809 Z = −0.589 
 padj = 0.653 padj = 0.546 padj = 0.013 padj = 0.014 padj = 0.020 padj = 0.556 
 Squelch 
  S0NSSD −9.6 −7.5 −8.8 −10.5 χ2(3) = 18.5 −2.9 (−6.9–2.3) −0.4 (−7.9–2.8) 2.2 (−3.6–4.2) 4.0 (1.9–6.7) 1.8 (−1.5–7.8) −1.5 (−5.4–3.3) 
p < 0.001 Z = −2.589 Z = −0.746 Z = −1.883 Z = −3.059 Z = −2.669 Z = −2.001 
 padj = 0.039 padj = 0.456 padj = 0.119 padj = 0.013 padj = 0.038 padj = 0.136 
  SAHNSSD −13.6 −10.8 −12.3 −14.6 χ2(3) = 24.9 −3.2 (−4.8–1.1) −1.4 (−5.8–2.0) 0.6 (−1.6–4.2) 3.4 (1.4–6.7) 1.4 (0.2–6.5) −1.9 (−3.4–2.2) 
p < 0.001 Z = −2.903 Z = −1.884 Z = −1.883 Z = −3.061 Z = −3.061 Z = −2.312 
 padj = 0.015 padj = 0.119 padj = 0.060 padj = 0.013 padj = 0.013 padj = 0.062 

Differences in speech reception threshold (SRT) between device conditions and preoperatively unaided represent binaural effects, and SRT differences between device conditions correspond to differences in binaural effects between devices. Positive SRT differences indicate binaural benefits in case of comparisons to preoperatively unaided, and differences in effects in case of comparisons between devices, respectively. Significant differences are highlighted in bold. Note that the median differences as reported here do not necessarily correspond to the differences of the medians.

AH, acoustic hearing; SNR, signal-to-noise ratio; SSD, single-sided deafness.

In addition, binaural spatial release from masking was evaluated as differences in SRT between the S0N0 and both the S0NAH and S0NSSD configurations for all masker and device conditions. The effect of device condition on SRM was analyzed independently for each masker and SRM configuration. Results from the respective Friedman tests and post hoc multiple comparisons are presented in Table 3. Figure 3 presents the binaural effects of head shadow, summation, squelch, and SRM for both maskers and the three device conditions.

Table 3.

Binaural spatial release from masking (SRM) for speech understanding in two-talker babble (TTB) and speech-shaped noise (SSN) within and between devices (CROS and BCD trials, CI at 6 months)

Median SRM, dBFriedman testMedian and range of SRM difference, dB
CROSBCDCI 6mCI 6m versus CROSCI 6m versus BCDCROS versus BCD
TTB 
 SRM (S0N0–S0NSSD1.4 3.0 6.2 χ2(2) = 11.2 4.8 (−1.4–8.2) 1.9 (−0.5–7.1) −0.8 (−5.6–2.4) 
Z = 2.446 Z = 2.982 Z = 3.061 p = 0.004 Z = −2.864 Z = −2.824 Z = −1.531 
p = 0.014 p = 0.003 p = 0.002  padj = 0.013 padj = 0.009 padj = 0.126 
 SRM (S0N0–S0NAH−2.0 −1.6 0.4 χ2(2) = 6.2 3.7 (−3.2–8.9) 2.1 (−3.5–10.4) 0.0 (−3.9–3.3) 
Z = −1.805 Z = −1.726 Z = 1.767 p = 0.046 Z = −2.197 Z = −2.158 Z = −0.589 
p = 0.071 p = 0.084 p = 0.077  padj = 0.084 padj = 0.062 padj = 0.556 
SSN 
 SRM (S0N0–S0NSSD1.9 3.9 4.9 χ2(2) = 12.2 2.7 (0.2–5.5) 0.4 (−1.5–6.3) −2.4 (−4.0–1.4) 
Z = 2.981 Z = 2.983 Z = 3.061 p = 0.002 Z = −3.061 Z = −1.804 Z = −1.961 
p = 0.003 p = 0.003 p = 0.002  padj = 0.007 padj = 0.071 padj = 0.100 
 SRM (S0N0–S0NAH−2.6 −2.5 0.0 χ2(2) = 15.5 3.1 (0.0–6.9) 2.1 (−2.2–7.2) −0.5 (−4.4–4.1) 
Z = −3.061 Z = −2.825 Z = 0.392 p < 0.001 Z = −3.062 Z = −2.590 Z = −0.235 
p = 0.002 p = 0.005 p = 0.695  padj = 0.007 padj = 0.019 padj = 0.814 
Median SRM, dBFriedman testMedian and range of SRM difference, dB
CROSBCDCI 6mCI 6m versus CROSCI 6m versus BCDCROS versus BCD
TTB 
 SRM (S0N0–S0NSSD1.4 3.0 6.2 χ2(2) = 11.2 4.8 (−1.4–8.2) 1.9 (−0.5–7.1) −0.8 (−5.6–2.4) 
Z = 2.446 Z = 2.982 Z = 3.061 p = 0.004 Z = −2.864 Z = −2.824 Z = −1.531 
p = 0.014 p = 0.003 p = 0.002  padj = 0.013 padj = 0.009 padj = 0.126 
 SRM (S0N0–S0NAH−2.0 −1.6 0.4 χ2(2) = 6.2 3.7 (−3.2–8.9) 2.1 (−3.5–10.4) 0.0 (−3.9–3.3) 
Z = −1.805 Z = −1.726 Z = 1.767 p = 0.046 Z = −2.197 Z = −2.158 Z = −0.589 
p = 0.071 p = 0.084 p = 0.077  padj = 0.084 padj = 0.062 padj = 0.556 
SSN 
 SRM (S0N0–S0NSSD1.9 3.9 4.9 χ2(2) = 12.2 2.7 (0.2–5.5) 0.4 (−1.5–6.3) −2.4 (−4.0–1.4) 
Z = 2.981 Z = 2.983 Z = 3.061 p = 0.002 Z = −3.061 Z = −1.804 Z = −1.961 
p = 0.003 p = 0.003 p = 0.002  padj = 0.007 padj = 0.071 padj = 0.100 
 SRM (S0N0–S0NAH−2.6 −2.5 0.0 χ2(2) = 15.5 3.1 (0.0–6.9) 2.1 (−2.2–7.2) −0.5 (−4.4–4.1) 
Z = −3.061 Z = −2.825 Z = 0.392 p < 0.001 Z = −3.062 Z = −2.590 Z = −0.235 
p = 0.002 p = 0.005 p = 0.695  padj = 0.007 padj = 0.019 padj = 0.814 

SRM effects are presented as differences in SRT between the S0N0 and both the S0NAH and S0NSSD configurations, respectively. Positive differences indicate a larger SRM for the first versus the second device condition, and significant differences are highlighted in bold. Note that the median differences as reported here do not necessarily correspond to the differences of the medians.

Fig. 3.

Box and whisker plots of binaural effects of head shadow, summation, squelch, and spatial release from masking (SRM) for two maskers. Top panels include two-talker babble (TTB) and bottom panels include speech-shaped noise (SSN). Head shadow (SSSDNAH, S0NAH), squelch (S0NSSD, SAHNSSD), and SRM (S0N0–S0NSSD and S0N0–S0NAH) were assessed for two spatial configurations each. Binaural effects were obtained for each device condition. The boxes show interquartile range (IQR), the whiskers show the data range excluding outliers, where outliers are defined as values further out than 1.5 times IQR from the outer box edges. Solid lines within the boxes represent the median and diamonds represent the mean. Significant binaural effects are denoted by asterisks. Brackets and asterisks represent significant differences between device conditions. *p < 0.05.

Fig. 3.

Box and whisker plots of binaural effects of head shadow, summation, squelch, and spatial release from masking (SRM) for two maskers. Top panels include two-talker babble (TTB) and bottom panels include speech-shaped noise (SSN). Head shadow (SSSDNAH, S0NAH), squelch (S0NSSD, SAHNSSD), and SRM (S0N0–S0NSSD and S0N0–S0NAH) were assessed for two spatial configurations each. Binaural effects were obtained for each device condition. The boxes show interquartile range (IQR), the whiskers show the data range excluding outliers, where outliers are defined as values further out than 1.5 times IQR from the outer box edges. Solid lines within the boxes represent the median and diamonds represent the mean. Significant binaural effects are denoted by asterisks. Brackets and asterisks represent significant differences between device conditions. *p < 0.05.

Close modal

Two-Talker Babble. For the TTB masker, Friedman tests revealed a significant effect of treatment condition on SRT in all spatial configurations (Table 2).

In the SSSDNAH configuration, significant improvements for speech understanding in noise compared to preoperatively unaided (i.e., significant head shadow benefits) were observed with CI at 6 months (7.8 dB) and with CROS (2.4 dB). No significant head shadow effect was found with BCD. Across devices, CI at 6 months showed a significantly larger head shadow effect compared to both CROS (by 5.8 dB) and BCD (by 6.2 dB). Compared to BCD, CROS showed no significant difference in SRT, and thus no significant difference in head shadow. In S0NAH, we found a significant head shadow benefit compared to unaided only with CI at 6 months (5.1 dB). Neither CROS nor BCD showed a significant improvement in SRT for this configuration. CI at 6 months showed a significantly larger head shadow effect than both CROS (by 4.9 dB) and BCD (by 4.2 dB). CROS showed no significant difference compared to BCD.

In S0N0, we observed significant improvements in SRT compared to unaided (i.e., significant summation benefits) for both CI at 6 months (3.4 dB) and BCD (1.5 dB). However, we found no significant summation effect with CROS. CI at 6 months yielded a significantly larger summation effect versus CROS (by 1.9 dB) and BCD (by 1.6 dB). CROS and BCD did not differ significantly.

In S0NSSD, there was a significant improvement in SRT compared to unaided (i.e., a significant squelch benefit) with CI at 6 months (4.6 dB). CROS use resulted in a significant deterioration in SRT (2.5 dB) (i.e., a significant negative squelch effect) while BCD use resulted in no significant difference. CI at 6 months showed a significantly larger squelch effect compared to both CROS (by 6.7 dB) and BCD (by 4.1 dB). CROS and BCD did not differ significantly. In SAHNSSD, we found a significant improvement in SRT compared to unaided only with CI at 6 months (5.1 dB). However, a significant negative squelch was found with CROS (−2.5 dB), while BCD showed no significant difference. CI at 6 months showed a significantly larger squelch effect than both CROS (by 8.6 dB) and BCD (by 6.3 dB), while there was no significant difference between CROS and BCD.

For TTB, Friedman tests revealed a significant effect of device condition on SRM in both SRM configurations (Table 3). Positioning the masker at the deaf side (S0NSSD) compared to the front (S0N0), we found a significant benefit in SRM of 6.2 dB with CI at 6 months, 3.0 dB with BCD, and 1.4 dB with CROS. CI at 6 months demonstrated a significantly larger SRM compared to both CROS (by 4.8 dB) and BCD (by 1.9 dB). SRM was not significantly different between CROS and BCD. Comparing the masker position at the acoustic-hearing side (S0NAH) to S0N0, SRM was not significant for any of the devices, nor did SRM differ significantly between devices.

Speech-Shaped Noise. Same as for TTB, Friedman tests showed a significant effect of treatment condition on SRT in each spatial configuration for the SSN masker (Table 2).

In SSSDNAH, we found significant improvements in speech understanding in noise compared to preoperatively unaided (i.e., significant head shadow benefits) with CI at 6 months (8.5 dB) and with CROS (1.7 dB). However, BCD showed no significant head shadow effect. CI at 6 months showed a significantly larger head shadow effect than both CROS (by 6.6 dB) and BCD (by 8.1 dB). Further, CROS and BCD did not differ significantly in head shadow. We observed a significant improvement in SRT compared to unaided in S0NAH only with CI at 6 months (4.6 dB). No significant head shadow effect was found for either CROS or BCD. CI at 6 months showed a significantly larger head shadow effect compared to both CROS (by 3.8 dB) and BCD (by 3.9 dB). We found no significant difference between CROS and BCD.

In S0N0, a significant improvement in SRT compared to unaided (i.e., a significant summation benefit) was only found with CI at 6 months (1.4 dB). Neither CROS nor BCD showed a significant summation effect. CI at 6 months demonstrated a significantly larger summation effect compared to both CROS (by 1.4 dB) and BCD (by 0.9 dB). Further, CROS showed no significant difference compared to BCD.

In S0NSSD, we found no significant difference in SRT compared to unaided (no significant squelch effect) with CI at 6 months and BCD. A significant negative squelch effect (−2.9 dB) was observed with CROS. CI at 6 months showed a significantly larger squelch effect than both CROS (by 4.0 dB) and BCD (by 1.8 dB). We found no significant difference between CROS and BCD. In SAHNSSD, no significant difference in SRT compared to unaided was observed with CI at 6 months or BCD. CROS resulted in a significant negative squelch effect (−3.2 dB). CI at 6 months showed a significantly larger squelch effect compared to both CROS (by 3.4 dB) and BCD (by 1.4 dB). CROS showed no significant difference to BCD.

For SSN and consistent with TTB, Friedman tests revealed a significant effect of device condition on SRM in both configurations (Table 3). Positioning the masker at the deaf side (S0NSSD) compared to the front (S0N0) resulted in a significant SRM benefit of 4.9 dB with CI at 6 months, 3.9 dB with BCD, and 1.9 dB with CROS. CI at 6 months yielded a significantly larger SRM than CROS (by 2.7 dB), but there was no significant difference to BCD. Further, CROS and BCD were not significantly different. With the masker positioned at the acoustic-hearing side (S0NAH), SRM was not significant with CI at 6 months. However, with CROS and BCD, SRM was negative and significant (−2.6 dB, −2.5 dB). CI at 6 months supported a significantly larger SRM compared to both CROS (by 3.1 dB) and BCD (by 2.1 dB), while SRM was not significantly different between CROS and BCD.

Effect of Masker. In the preoperative unaided condition, the ANOVA revealed no significant effect of masker (F [1, 11] = 0.254, p = 0.624), a significant effect of spatial configuration (F [1.921, 21.133] = 117.747, p < 0.001), and no significant masker × spatial configuration interaction on SRT (F [4, 44] = 2.311, p = 0.073).

With CI at 6 months, there was a significant effect of masker (F [1, 11] = 11.607, p = 0.006), a significant effect of spatial configuration (F [2.298, 25.279] = 48.792, p < 0.001), and no significant masker × spatial configuration interaction on SRT (F [1.476, 16.233] = 1.535, p = 0.242). Across spatial configurations, SRT in TTB (−10.6 dB SNR) was 1.5 dB better compared to SSN (−9.1 dB SNR). Pairwise post hoc tests revealed significant differences in SRT in S0NAH (1.9 dB better in TTB, padj = 0.026) and S0NSSD (2.0 dB better in TTB, padj < 0.001).

For the binaural effects comparing CI at 6 months with preoperatively unaided, there was a significant effect of masker (F [1, 11] = 26.825, p < 0.001), a significant effect of spatial configuration (F [4, 44] = 19.094, p < 0.001), and no significant masker × spatial configuration interaction (F [1.970, 21.666] = 2.595, p = 0.098). Binaural benefit in TTB (5.1 dB) was 1.8 dB larger compared to SSN (3.3 dB) across spatial configurations. Pairwise post hoc tests revealed significant differences in binaural effect in S0N0 (2.2 dB better in TTB, padj < 0.001) and S0NSSD (2.4 dB better in TTB, padj = 0.020).

Effect of CI Use over Time

For the three post-activation CI intervals, speech understanding in noise was analyzed independently for the five spatial configurations and both maskers. Table 4 presents results from the respective Friedman tests and post hoc multiple comparisons.

Table 4.

Longitudinal speech understanding with CI in two-talker babble (TTB) and speech-shaped noise (SSN) across spatial configurations

Median SRT, dB SNRFriedman testMedian and range of SRT difference, dB
preoperatively unaidedCI 1mCI 3mCI 6mCI 1m versus preoperatively unaidedCI 3m versus preoperatively unaidedCI 6m versus preoperatively unaidedCI 3m versus CI 1mCI 6m versus CI 3mCI 6m versus CI 1m
TTB 
 Head shadow 
  SSSDNAH 0.2 −5.8 −7.1 −8.2 χ2(3) = 20.8 5.8 (−2.5–15.9) 7.3 (−2.0–12.1) 7.8 (2.4–12.0) 2.2 (−6.8–9.7) 0.9 (−1.4–4.4) 3.2 (−4.0–10.4) 
p < 0.001 Z = −2.824 Z = −2.981 Z = −3.061 Z = −1.098 Z = −2.040 Z = −1.806 
 padj = 0.019 padj = 0.014 padj = 0.013 padj = 0.272 padj = 0.124 padj = 0.142 
  S0NAH −3.3 −3.4 −6.3 −8.0 χ2(3) = 24.3 3.0 (−1.9–7.0) 3.4 (−3.6–8.1) 5.1 (1.3–11.3) 1.5 (−4.3–4.2) 2.2 (−1.4–7.4) 2.9 (0.3–7.8) 
p < 0.001 Z = −2.119 Z = −2.589 Z = −3.059 Z = −1.647 Z = −2.630 Z = −3.059 
 padj = 0.068 padj = 0.029 padj = 0.013 padj = 0.099 padj = 0.034 padj = 0.013 
 Summation 
  S0N0 −3.6 −7.3 −7.6 −7.3 χ2(3) = 21.9 3.9 (1.1–4.6) 3.1 (1.8–6.0) 3.4 (2.9–5.3) 0.0 (−2.5–1.9) 0.3 (−2.7–1.7) −0.4 (−1.3–2.1) 
p < 0.001 Z = −3.061 Z = −3.061 Z = −3.061 Z = 0.000 Z = −0.118 Z = −0.471 
 padj = 0.013 padj = 0.013 padj = 0.013 padj = 1 padj = 1 padj = 1 
 Squelch 
  S0NSSD −9.0 −11.7 −11.6 −12.6 χ2(3) = 13.7 2.4 (−7.0–6.0) 3.4 (−5.4–6.2) 4.6 (−1.8–9.0) 0.3 (−6.1–8.9) 0.7 (−3.7–6.6) 0.9 (−0.7–8.4) 
p = 0.003 Z = −1.766 Z = −1.961 Z = −2.667 Z = −0.707 Z = −1.886 Z = −2.667 
 padj = 0.155 padj = 0.199 padj = 0.046 padj = 0.479 padj = 0.178 padj = 0.046 
  SAHNSSD −13.6 −16.2 −16.4 −17.7 χ2(3) = 12.9 3.6 (−4.4–7.9) 4.2 (−6.2–14.4) 5.1 (−3.6–10.6) 1.2 (−11.9–7.0) 0.8 (−7.9–16.8) 1.3 (−7.1–5.5) 
p = 0.005 Z = −2.510 Z = −1.961 Z = −2.432 Z = −0.628 Z = −0.314 Z = −0.941 
 padj = 0.072 padj = 0.199 padj = 0.075 padj = 1 padj = 0.754 padj = 0.347 
SSN 
 Head shadow 
  SSSDNAH 1.9 −4.3 −6.6 −8.2 χ2(3) = 27.5 5.3 (−1.1–11.6) 7.6 (1.9–12.7) 8.5 (3.3–13.0) 2.0 (−5.5–9.0) 1.7 (−7.0–5.5) 4.4 (−5.6–10.4) 
p < 0.001 Z = −2.904 Z = −3.059 Z = −3.059 Z = −2.353 Z = −2.002 Z = −2.315 
 padj = 0.015 padj = 0.013 padj = 0.013 padj = 0.056 padj = 0.045 padj = 0.041 
  S0NAH −1.1 −3.2 −5.7 −5.9 χ2(3) = 25.3 2.4 (−2.3–4.1) 4.1 (0.8–9.4) 4.6 (1.0–7.5) 1.8 (−2.0–7.1) 0.3 (−3.1–6.8) 2.7 (0.1–6.9) 
p < 0.001 Z = −2.040 Z = −3.061 Z = −3.061 Z = −2.353 Z = −0.784 Z = −3.059 
 padj = 0.083 padj = 0.013 padj = 0.013 padj = 0.056 padj = 0.433 padj = 0.009 
 Summation 
  S0N0 −4.7 −5.7 −5.6 −5.8 χ2(3) = 25.6 0.8 (−0.3–2.4) 1.4 (0.1–2.3) 1.4 (0.2–2.8) 0.6 (−1.1–1.3) 0.1 (−0.7–0.7) 0.4 (−0.9–1.8) 
p < 0.001 Z = −2.824 Z = −3.062 Z = −3.062 Z = −1.735 Z = −1.415 Z = −1.961 
 padj = 0.019 padj = 0.013 padj = 0.013 padj = 0.166 padj = 0.157 padj = 0.150 
 Squelch 
  S0NSSD −9.6 −8.9 −9.3 −10.5 χ2(3) = 11.5 0.2 (−5.0–2.6) 0.7 (−3.3–3.4) 2.2 (−3.6–4.2) 0.6 (−3.6–6.7) 1.1 (−0.6–3.6) 1.5 (−1.4–8.3) 
p = 0.009 Z = −0.275 Z = −0.981 Z = −1.883 Z = −1.413 Z = −2.395 Z = −2.747 
 padj = 0.784 padj = 0.653 padj = 0.239 padj = 0.473 padj = 0.083 padj = 0.036 
  SAHNSSD −13.6 −14.8 −13.9 −14.6 χ2(3) = 3.3 0.7 (−4.9–6.1) 1.0 (−4.7–9.9) 0.6 (−1.6–4.2) 0.0 (−5.0–6.1) −0.1 (−5.7–5.0) 0.1 (−2.0–7.2) 
p = 0.348 Z = −1.491 Z = −1.020 Z = −1.883 Z = −0.471 Z = −0.078 Z = −0.235 
 padj = 0.680 padj = 1 padj = 0.358 padj = 1 padj = 0.937 padj = 1 
Median SRT, dB SNRFriedman testMedian and range of SRT difference, dB
preoperatively unaidedCI 1mCI 3mCI 6mCI 1m versus preoperatively unaidedCI 3m versus preoperatively unaidedCI 6m versus preoperatively unaidedCI 3m versus CI 1mCI 6m versus CI 3mCI 6m versus CI 1m
TTB 
 Head shadow 
  SSSDNAH 0.2 −5.8 −7.1 −8.2 χ2(3) = 20.8 5.8 (−2.5–15.9) 7.3 (−2.0–12.1) 7.8 (2.4–12.0) 2.2 (−6.8–9.7) 0.9 (−1.4–4.4) 3.2 (−4.0–10.4) 
p < 0.001 Z = −2.824 Z = −2.981 Z = −3.061 Z = −1.098 Z = −2.040 Z = −1.806 
 padj = 0.019 padj = 0.014 padj = 0.013 padj = 0.272 padj = 0.124 padj = 0.142 
  S0NAH −3.3 −3.4 −6.3 −8.0 χ2(3) = 24.3 3.0 (−1.9–7.0) 3.4 (−3.6–8.1) 5.1 (1.3–11.3) 1.5 (−4.3–4.2) 2.2 (−1.4–7.4) 2.9 (0.3–7.8) 
p < 0.001 Z = −2.119 Z = −2.589 Z = −3.059 Z = −1.647 Z = −2.630 Z = −3.059 
 padj = 0.068 padj = 0.029 padj = 0.013 padj = 0.099 padj = 0.034 padj = 0.013 
 Summation 
  S0N0 −3.6 −7.3 −7.6 −7.3 χ2(3) = 21.9 3.9 (1.1–4.6) 3.1 (1.8–6.0) 3.4 (2.9–5.3) 0.0 (−2.5–1.9) 0.3 (−2.7–1.7) −0.4 (−1.3–2.1) 
p < 0.001 Z = −3.061 Z = −3.061 Z = −3.061 Z = 0.000 Z = −0.118 Z = −0.471 
 padj = 0.013 padj = 0.013 padj = 0.013 padj = 1 padj = 1 padj = 1 
 Squelch 
  S0NSSD −9.0 −11.7 −11.6 −12.6 χ2(3) = 13.7 2.4 (−7.0–6.0) 3.4 (−5.4–6.2) 4.6 (−1.8–9.0) 0.3 (−6.1–8.9) 0.7 (−3.7–6.6) 0.9 (−0.7–8.4) 
p = 0.003 Z = −1.766 Z = −1.961 Z = −2.667 Z = −0.707 Z = −1.886 Z = −2.667 
 padj = 0.155 padj = 0.199 padj = 0.046 padj = 0.479 padj = 0.178 padj = 0.046 
  SAHNSSD −13.6 −16.2 −16.4 −17.7 χ2(3) = 12.9 3.6 (−4.4–7.9) 4.2 (−6.2–14.4) 5.1 (−3.6–10.6) 1.2 (−11.9–7.0) 0.8 (−7.9–16.8) 1.3 (−7.1–5.5) 
p = 0.005 Z = −2.510 Z = −1.961 Z = −2.432 Z = −0.628 Z = −0.314 Z = −0.941 
 padj = 0.072 padj = 0.199 padj = 0.075 padj = 1 padj = 0.754 padj = 0.347 
SSN 
 Head shadow 
  SSSDNAH 1.9 −4.3 −6.6 −8.2 χ2(3) = 27.5 5.3 (−1.1–11.6) 7.6 (1.9–12.7) 8.5 (3.3–13.0) 2.0 (−5.5–9.0) 1.7 (−7.0–5.5) 4.4 (−5.6–10.4) 
p < 0.001 Z = −2.904 Z = −3.059 Z = −3.059 Z = −2.353 Z = −2.002 Z = −2.315 
 padj = 0.015 padj = 0.013 padj = 0.013 padj = 0.056 padj = 0.045 padj = 0.041 
  S0NAH −1.1 −3.2 −5.7 −5.9 χ2(3) = 25.3 2.4 (−2.3–4.1) 4.1 (0.8–9.4) 4.6 (1.0–7.5) 1.8 (−2.0–7.1) 0.3 (−3.1–6.8) 2.7 (0.1–6.9) 
p < 0.001 Z = −2.040 Z = −3.061 Z = −3.061 Z = −2.353 Z = −0.784 Z = −3.059 
 padj = 0.083 padj = 0.013 padj = 0.013 padj = 0.056 padj = 0.433 padj = 0.009 
 Summation 
  S0N0 −4.7 −5.7 −5.6 −5.8 χ2(3) = 25.6 0.8 (−0.3–2.4) 1.4 (0.1–2.3) 1.4 (0.2–2.8) 0.6 (−1.1–1.3) 0.1 (−0.7–0.7) 0.4 (−0.9–1.8) 
p < 0.001 Z = −2.824 Z = −3.062 Z = −3.062 Z = −1.735 Z = −1.415 Z = −1.961 
 padj = 0.019 padj = 0.013 padj = 0.013 padj = 0.166 padj = 0.157 padj = 0.150 
 Squelch 
  S0NSSD −9.6 −8.9 −9.3 −10.5 χ2(3) = 11.5 0.2 (−5.0–2.6) 0.7 (−3.3–3.4) 2.2 (−3.6–4.2) 0.6 (−3.6–6.7) 1.1 (−0.6–3.6) 1.5 (−1.4–8.3) 
p = 0.009 Z = −0.275 Z = −0.981 Z = −1.883 Z = −1.413 Z = −2.395 Z = −2.747 
 padj = 0.784 padj = 0.653 padj = 0.239 padj = 0.473 padj = 0.083 padj = 0.036 
  SAHNSSD −13.6 −14.8 −13.9 −14.6 χ2(3) = 3.3 0.7 (−4.9–6.1) 1.0 (−4.7–9.9) 0.6 (−1.6–4.2) 0.0 (−5.0–6.1) −0.1 (−5.7–5.0) 0.1 (−2.0–7.2) 
p = 0.348 Z = −1.491 Z = −1.020 Z = −1.883 Z = −0.471 Z = −0.078 Z = −0.235 
 padj = 0.680 padj = 1 padj = 0.358 padj = 1 padj = 0.937 padj = 1 

Speech reception thresholds (SRTs) for each test interval and comparisons in SRT between intervals are shown. Positive SRT differences indicate better performance for the first versus the second interval, and significant differences are highlighted in bold. Note that the median differences as reported here do not necessarily correspond to the differences of the medians.

SNR, signal-to-noise ratio.

Two-Talker Babble. For the TTB masker, Friedman tests revealed a significant effect of test interval on SRT in all spatial configurations (Table 4). In SSSDNAH, we observed significant improvements in SRT compared to preoperatively unaided (i.e., significant head shadow benefits) with CI at all intervals of 1 (5.8 dB), 3 (7.3 dB), and 6 months (7.8 dB). However, the SRTs with CI did not differ significantly between the intervals. In S0NAH, SRTs improved significantly with CI at 3 (3.4 dB) and 6 months (5.1 dB) compared to unaided, but not at 1 month. CI performance improved significantly from 1 to 6 months (2.9 dB) and from 3 to 6 months (2.2 dB); however, there was no significant difference in SRT between 1 and 3 months.

We found significant improvements in SRT with CI at 1 (3.9 dB), 3 (3.1 dB), and 6 months (3.4 dB) compared to unaided (i.e., significant summation benefits) in S0N0. Further, the postoperative intervals did not significantly differ.

In S0NSSD, CI usage led to a significant squelch benefit at 6 months with an improvement in SRT of 4.6 dB compared to unaided, but neither at 1 nor 3 months. However, there was a significant improvement in SRT from 1 to 6 months (0.9 dB), but neither between 1 and 3 nor between 3 and 6 months. In SAHNSSD, CI did not show a significant difference in SRT compared to unaided at any interval. This includes the 6-month interval for which, in contrast, a significant squelch effect was shown in the previous analysis of effect of treatment due to different p value ranks in the Bonferroni-Holm correction. SRTs with CI did not differ significantly between the intervals.

Speech-Shaped Noise. For the SSN masker, Friedman tests revealed a significant effect of test interval on SRT in all spatial configurations except SAHNSSD (Table 4). In SSSDNAH, SRTs with CI improved significantly at all intervals including 1 (5.3 dB), 3 (7.6 dB), and 6 months (8.5 dB) compared to preoperatively unaided (i.e., significant head shadow benefits). The SRT improvements with CI were significant from 1 to 6 (4.4 dB) and from 3 to 6 months (1.7 dB), however, SRT did not change significantly from 1 to 3 months. Significant improvements in SRT compared to unaided were observed in S0NAH with CI at 3 (4.1 dB) and 6 months (4.6 dB), but not at 1 month. CI performance improved significantly from 1 to 6 months (2.7 dB), however, did neither change significantly from 1 to 3 nor from 3 to 6 months.

We found significantly better SRTs with CI at 1 (0.8 dB), 3 (1.4 dB), and 6 months (1.4 dB) compared to unaided in S0N0 (i.e., significant summation benefits). However, there were no significant differences between the postoperative intervals.

Compared to preoperatively unaided, SRT with CI did not change significantly at any interval in S0NSSD (i.e., no significant squelch effect). However, SRT improved significantly from 1 to 6 months (1.5 dB), but was not significantly different between the other intervals. In SAHNSSD, no significant difference in SRT with CI compared to unaided was observed at any interval, nor were there significant differences between the postoperative intervals.

The primary aim of this study was to investigate masked speech understanding and binaural benefits in people with SSD comparing preoperative CROS and BCD as well as CI to preoperative unaided baseline and between devices. Furthermore, we assessed the effect of masker type on speech understanding. The study design followed recommendations from the consensus protocol [5].

The two maskers for evaluating speech perception were OLnoise, a SSN which acted as a purely energetic masker, and a TTB which acted also as an informational masker. Testing was carried out in five spatial speech-in-noise configurations (SSSDNAH, S0NAH, S0N0, S0NSSD, SAHNSSD) to assess binaural effects of head shadow, summation, squelch, and SRM.

Twelve participants completed the preoperative trials with CROS and BCD, and all twelve opted to receive a CI. Their performance with a CI was evaluated at 1, 3, and 6 months post-activation. Results will be herein discussed and compared with previous studies in terms of observed binaural effects, comparisons among the different treatment devices, the effects of the different masker types, and longitudinal effects of CI usage. Binaural effects are referred to as differences in aided and preoperatively unaided SRTs for head shadow, summation, and squelch.

Binaural Effects with CROS/BCD Trial Devices and CI

Head Shadow

To assess the head shadow effect, we used two different spatial configurations, SSSDNAH and S0NAH. Compared to preoperatively unaided, 6 months of CI usage led to the largest and most consistent improvements in SRT across head shadow configurations (SSSDNAH, S0NAH) and maskers compared to CROS and BCD. The CI supported significant head shadow benefits in all maskers and configurations. Whereas CROS only yielded significant benefits in the SSSDNAH configuration (i.e., the “wider” of the two head shadow configurations). With BCD, we found no significant change in any configuration or masker. Comparing devices, head shadow effects with CROS were of considerably and significantly smaller magnitude than those with CI across maskers and configurations. Similarly, CI also supported substantially, and significantly larger head shadow effects compared to BCD in all maskers and configurations. There were no significant differences in head shadow between CROS and BCD for any configuration and masker.

In different SSSDNAH configurations, significant head shadow benefits gained with CI in recipients with SSD compared to preoperatively unaided have been previously reported. When using an S±45°N∓45° configuration with less spatial separation between speech and masker, after 12 months of CI experience, Arndt et al. [9] found a significant improvement in SRT (7.1 dB) for OLSA sentences in OLnoise. The larger head shadow effect of 8.5 dB found in the present study might be explained at least in part by the wider spatial separation between target and masker of ±90° in our SSSDNAH configuration. This is supported by the results of Kurz et al. [7] who also observed an SRT improvement of 8.5 dB with CI for OLSA sentences in OLnoise at 6 months and for the same SSSDNAH configuration as in our study. Further, Peters et al. [25] reported a significant improvement of 6.9 dB for Utrecht sentences in SSN using an S±60°N∓60° configuration with CI at 6 months; and Jakob et al. [11] showed a significant increase in word score of approximately 44 percentage points for Hochmair-Schulz-Moser (HSM) sentences [26] in CCITT noise at an SNR of 0 dB in an S±45°N∓45° configuration after 12 months.

Similarly, Deep et al. [27] used an S0N±90° configuration and compared CI to preoperatively unaided. They tested speech performance for Bamford-Kowal-Bench (BKB) sentences in 30 SSD CI recipients and for Hearing in Noise Test (HINT) sentences in seven recipients, both in four-talker babble, at a mean follow-up of 2.2 years. However, albeit similar testing methodology, they observed a significant head shadow benefit of 2.6 dB, which is smaller than the 5.1 dB benefit observed in the present study. This might be due to the different masker used by Deep et al. [27]. Further, Buss et al. [28] measured speech understanding for AzBio sentences in 10-talker babble in an S0N±90° configuration and reported an improvement of 36 percentage points with a CI at 12 months compared to preoperatively unaided. Using SSN as masker, Galvin et al. [17] assessed speech in noise for HINT sentences in S0N±90° and observed an improvement of about 3.0 dB with CI at 6 months compared to preoperatively unaided, versus 4.6 dB in our study.

Arndt et al. [9] investigated head shadow with both CROS and BCD trial devices in an S±45°N∓45° configuration. They observed a significant head shadow benefit for OLSA sentences in OLnoise of about 2.9 dB with CROS only compared to 1.7 dB in the present study. However, as in our study, they found no significant effect with BCD. To address head shadow in the same configuration with CROS and BCD trial devices in a group later opting for a CI, Jakob et al. [11] assessed word scores for HSM sentences in CCITT noise at 0 dB SNR and found similar results, with significant benefits of using CROS of about 19 percentage points, and a smaller benefit of 5 percentage points with BCD. However, Kurz et al. [7], who tested SRTs for OLSA sentences in OLnoise in an S±90°N∓90° configuration, found a significant improvement in SRT of 3 dB with a BCD trial device, but no significant difference between CROS and preoperatively unaided. The heterogeneity of these findings might be related to different spatial configurations and different types of CROS and BCD trial devices. Using an S0N±90° configuration, Deep et al. [27] found no significant difference in SRT for BKB or HINT sentences in four-talker babble with a CROS trial device compared to preoperatively unaided. This is consistent with our findings with CROS for OLSA sentences when using the TTB masker.

Both Arndt et al. [9] and Kurz et al. [7] compared SRTs in SSN between devices in different SSSDNAH configurations. Using an S±45°N∓45° configuration, Arndt et al. [9] reported a significantly larger head shadow effect with CI at 12 months compared to both CROS and BCD trial devices by about 4 dB and 5.9 dB. The head shadow effect between CROS and BCD was not significantly different. The larger head shadow effect in SSN with CI compared to both CROS and BCD, in the current study, can be attributed to a better SRT with CI by approximately 1.8 dB and poorer SRTs with CROS and BCD by about 1.3 dB and 1.4 dB. In the same S±90°N∓90° configuration and consistent with our results, Kurz et al. [7] found a larger head shadow effect with CI at 6 months than with either CROS (by 3.4 dB) or BCD (by 3.6 dB) trial devices. Testing sentences in four-talker babble, Deep et al. [27] found a significantly larger head shadow effect by 3.6 dB with CI at a mean follow-up of 2.2 years versus CROS trial device in the S0N±90° configuration. A similar observation was made in our study with a slightly larger head shadow effect in TTB by 4.9 dB with CI compared to CROS.

Summation

In the summation configuration (S0N0) and compared to preoperatively unaided, CI supported the largest and most consistent improvements in SRT across devices and maskers. With CI at 6 months, we observed significant summation benefits in both maskers. With the trial devices, a significant summation benefit was observed only with BCD in TTB. With BCD in SSN and with CROS in both maskers, there was no significant summation effect. Across devices, summation effects with CI were significantly larger compared to both CROS and BCD. There was no significant difference in summation between CROS and BCD in either masker.

Using SSN as a masker, studies on SSD CI recipients have found significant improvements in SRT for summation of about 0.7 dB at 12 months [9] and of about 1.2 dB [25], respectively, of 3.4 dB [17] at 6 months compared to preoperatively unaided. The significant SRT improvement of 1.4 dB in our study is comparable to these results. While Buss et al. [28] also found a significant improvement of 9.7 percentage points with CI at 12 months compared to preoperatively unaided, Jakob et al. [11] observed no significant difference for summation.

Deep et al. [27] tested SRTs in four-talker babble and found no significant summation effect with CI at a mean follow-up of 2.2 years. In contrast to these findings, our results demonstrate a significant summation benefit of 3.4 dB for TTB with CI at 6 months. This difference might be explained by the larger number and the mixed gender of the competing talkers used by Deep et al. [27] compared to the same gender of target and competing talkers in our study.

Concerning the results with CROS and BCD trial devices compared to preoperatively unaided, no significant summation effect was observed in any of the previous studies, regardless of masker type [9, 11, 27]. In the current study, we found a small but significant summation benefit with BCD in TTB (1.5 dB) only. When comparing the devices, Arndt et al. [9] showed a significant summation benefit in SSN with CI at 12 months, but no significant effect with CROS and BCD trial devices. Summation effect with CI was significantly larger by approximately 1.1 dB compared to CROS. Neither CI versus CROS, nor CROS versus BCD were significantly different. Consistently, the present study revealed a significantly larger summation effect by 1.4 dB in SSN with CI at 6 months compared to CROS, and no significant difference between CROS and BCD. While Arndt et al. [9] did not find a significant difference between CI and BCD, we observed a significantly larger summation effect by 0.9 dB with CI compared to BCD. However, Deep et al. [27] found no significant difference in summation in a four-talker babble between CI at a mean follow-up of 2.2 years and CROS. In TTB, we observed a significantly larger summation effect by 1.9 dB with CI at 6 months compared to CROS, which might again be attributed to the different babble masker applied by Deep et al. [27].

Squelch

We applied two different spatial configurations, S0NSSD and SAHNSSD, to evaluate the squelch effect. Compared to preoperatively unaided, 6 months of CI use led to significantly better SRTs in both configurations in TTB, albeit no significant difference in SSN. Further, no change in SRT was observed with BCD in any squelch configuration or masker. CROS use led to a significant deterioration in SRT in all maskers and squelch configurations. Across devices, squelch effects with CI at 6 months were significantly larger compared to both CROS and BCD across maskers and configurations. There was no significant difference in squelch between CROS and BCD in any configuration or masker.

A recent study from Deep et al. [27], while using the same S0N±90° configuration as in the current study, found no significant change in SRT in four-talker babble with CI at a mean follow-up of 2.2 years. Similarly, Buss et al. [28] demonstrated that speech recognition in a 10-talker babble did not differ significantly between CI at 12 months and preoperatively unaided. Our conflicting observation of a significant squelch benefit of 4.6 dB in TTB might be attributed to the application of a same-gender masker comprising only two competing talkers in contrast to the mixed 10-talker babble used in Buss et al. [28] and the mixed four-talker babble applied in Deep et al. [27]. In a squelch configuration, presenting both target and masker to the acoustic-hearing ear and only the masker to the CI ear, Bernstein et al. [29] found a significant benefit, acutely, with CI for same-gender maskers but not in opposite-gender maskers. Assessing SRTs using an SSN masker in an S0N±90° configuration, Galvin et al. [17] did not find a significant difference between CI at 6 months and preoperatively unaided, which is similar to our findings. Furthermore, no squelch effect when using the SSN masker was also demonstrated by Bernstein et al. [29].

Studies have assessed speech understanding with CI compared to preoperatively unaided in different SAHNSSD configurations other than S±90°N∓90° as in the current investigation [9, 11, 25]. In an S±60°N∓60° configuration, Peters et al. [25] found a small but significant SRT benefit of 0.6 dB with 6 months of CI use for Utrecht sentences in SSN. Other studies assessed SRTs or word recognition scores in SSN in an S±45°N∓45° configuration and observed no significant difference with CI at 12 months [9, 11]. This mirrors the findings of the present study, in which no significant squelch effect was observed when SSN was used as masker. However, we found a significant squelch benefit of 5.1 dB in S±90°N∓90° in TTB consistent with the significant benefit by Bernstein et al. [29] for a comparable squelch configuration using a TTB masker.

When using the CROS trial device in the same S0N±90° configuration as in the present study, Deep et al. [27] found a significant decrement in SRT of 2.1 dB in four-talker babble compared to preoperatively unaided. This result compares to the decrement of 2.5 dB for squelch in TTB in the current study. Previous investigations have also compared speech understanding in SSN with both CROS and BCD trial devices to preoperatively unaided in an S±45°N∓45° configuration. With CROS, they found significant decrements in SRT of approximately 2.5 dB [9] and in word score of approximately 4.4 percentage points [11] but no significant difference with BCD in either study. Likewise, in our study CROS yielded a significant deterioration in SRT in SSN of 3.2 dB (i.e., a significant negative squelch effect) whereas BCD showed no significant difference.

Comparing the devices, Deep et al. [27] used the same S0N±90° configuration and found no significant difference in squelch effect between CI and CROS. In contrast, we observed a significantly larger squelch effect in TTB with CI versus CROS (by 6.7 dB). In an S±45°N∓45° squelch configuration, Arndt et al. [9] also observed a significantly larger squelch effect in SSN with CI at 12 months compared to both CROS (by 4.6 dB) and BCD (by 2.3 dB). Similarly, we found a significantly larger squelch effect in SSN with CI at 6 months compared to CROS (by 3.4 dB) and BCD (by 1.4 dB) in the S±90°N∓90° configuration. While squelch in the Arndt et al. [9] investigation was significantly poorer with CROS compared to BCD by approximately 2.3 dB, squelch in our study was not significantly different between these devices.

Spatial Release from Masking

SRM was assessed in two symmetric configurations by positioning the masker at either the deaf (S0NSSD) or acoustic-hearing (S0NAH) side at ±90° compared to the front (S0N0). In the S0N0–S0NSSD configuration and both maskers, the largest and significant benefits were observed with CI at 6 months, but significant benefits were also observed with CROS and BCD trial devices.

Several studies have investigated SRM with CI using the same S0N0–S0NSSD configuration. Müller and Lang-Roth [30] assessed speech recognition for OLSA sentences in same-talker masker at 0 dB SNR and found a significant SRM benefit of 41 percentage points with CI at a mean follow-up of 12 months. This finding is in agreement with our SRM result of 6.2 dB in TTB, albeit the two studies have important methodological differences to consider. Using SSN as masker and OLSA sentences as a target, Grossmann et al. [31] observed a significant SRM benefit of 7.2 dB with CI at a mean follow-up of 22.7 months compared to 4.9 dB with CI at 6 months in our study. Similarly, testing for SRM using LIST sentences in SSN, Mertens et al. [10] showed a significant improvement of 4.67 dB with CI at 6 months.

For SRM in the opposite configuration (S0N0-S0NAH), we did not find a significant effect with CI at 6 months in either TTB or SSN. Both CROS and BCD, however, led to a significant deterioration in SRT (i.e., a significant negative SRM) in SSN, while there was no effect in TTB.

Measuring SRM in the same S0N0-S0NAH configuration, Müller and Lang-Roth [30] found a significant benefit of 27 percentage points for sentences in same-talker masker with CI at 12 months. Using SSN as masker, Grossmann et al. [31] observed a significant SRM benefit of 1.4 dB with CI at a mean follow-up of 22.7 months. These findings contrast with our study, which might be explained by the longer time intervals at which SRM was assessed.

General Considerations on Treatment Devices and Binaural Effects

In general, we observed that only CI yielded consistent and significant benefits for each binaural effect. This includes summation and squelch, which are seen as manifestations of true binaural processing because they require integration of the input from both ears [32]. In individuals with SSD, only CI use reactivates binaural processing by reinstating auditory input to the deaf ear. On the contrary, CROS and BCD are monaural treatment solutions for SSD; they merely transmit sounds to the acoustically hearing ear. As such, they can compensate the head shadow for target signals at the side of the deaf ear. However, in the squelch configurations with presentation of the masker at the deaf ear, both CROS and BCD may even deteriorate speech understanding by transmitting the masker to the contralateral acoustically hearing ear.

Consistently, we found head shadow benefits with the CROS trial device, albeit of smaller magnitude than with CI. An explanation for the larger head shadow benefit with CI might be reactivation of the deaf ear, enabling listening with the ear facing the better SNR and potentially providing some redundancy for listening in a head shadow configuration. Head shadow benefits with CROS are potentially limited by monaural hearing with the acoustically hearing ear and the superposition of a wirelessly transmitted and delayed target signal from the contralateral deaf ear over the direct sound in that ear. For the BCD trial device, a necessary limitation of the testing conducted here is that this device was worn with a headband rather than as a percutaneous device. The transcutaneous sound coupling results in a high-frequency signal attenuation [33]. In addition, there is also a transcranial attenuation of the vibratory sound transmission, particularly at frequencies between 2.5 and 5.7 kHz, to the acoustically hearing ear [34]. Taken together with the fact that the BCD is also a monaural hearing solution, this may explain the unexpected lack of head shadow benefit with BCD that was observed in our study.

Regarding summation, the significant benefits found with CI in both maskers support the notion that the CI enables true binaural processing by utilizing binaural redundancy and loudness summation. Compared to CI, there was no significant summation effect with either CROS or BCD trial devices except a benefit with BCD in TTB which was significantly smaller compared to CI. This benefit might mirror monaural loudness summation.

For binaural squelch as the second true binaural effect, the CI revealed significant benefits in TTB suggesting true binaural processing based on binaural contrast and redundancy [35]. With CROS and BCD as monaural treatment solutions, however, the signal with the poorer SNR is captured from the contralateral deaf ear and added to the better SNR signal at the acoustically hearing ear. This explains the observed deterioration in squelch with CROS in both maskers. Furthermore, the delay between the direct and the transmitted signals might have contributed to this negative effect. The observation of no detrimental squelch effect with BCD on a headband may be attributed to the transcutaneous and transcranial attenuations of the poorer SNR signal, similar to the lack of benefit in the head shadow configuration.

According to Dieudonné and Francart [35], SRM is based on changes in monaural cues at both ears as well as on the arising of binaural interaural time and level difference cues by moving the masker from the front to either side, the deaf ear or the acoustically hearing ear. In the SRM configurations investigated in the current study, there were two changes in the two monaural cues: (1) an increase in SNR at the acoustically hearing ear and a decrease in SNR at the aided contralateral ear in the S0N0–S0NSSD configuration and (2) vice versa in S0N0–S0NAH, the opposite configuration. In the S0N0–S0NSSD configuration, the SRM benefits with both CROS and BCD are due to the SNR increase at the acoustically hearing ear. The larger SRM with CI in TTB compared to both CROS and BCD and in SSN compared to CROS might reflect an effective use of binaural cues in addition to the monaural better-SNR cue. In the S0N0–S0NAH configuration and in SSN, the deterioration in SRM with CROS and BCD can be attributed to the decrease in SNR at the acoustically hearing ear. In contrast, the CI was able to compensate this SNR decrement by presenting the better-SNR signal to the deaf ear, resulting in no deterioration in SRM overall.

The large majority of the observed statistically significant differences in binaural effects was greater than 1 dB and was thus considered to be clinically relevant [36, 37].

Effect of Masker

In this study, two different maskers were used: TTB and SSN. Because these maskers differ in both temporal and spectral properties as well as in semantic content, it is conceivable that differences exist in speech performance when either masker is applied.

With CI at 6 months, there was a significant effect of masker on SRT. Across spatial configurations, SRT in TTB was 1.5 dB better than in SSN. There are several potential explanations for why it is easier to understand speech in TTB. First, SSN is almost completely unmodulated, while TTB is characterized by pronounced temporal fluctuations. Such modulations might allow for gap listening. This is supported by the findings of Prejban et al. [38], who tested speech intelligibility in noise in two spatial configurations in SSD CI recipients at a mean follow-up of 14.3 months. They found an SRT improvement of 3.5 dB across spatial configurations in temporally fluctuating Fastl noise [39] compared to OLnoise, the same SSN masker as applied in the current study. Second, in our study, the target voice and the two babble voices were recorded from three different male speakers. The long-term spectra of these three voices do not completely overlap. The SSN, on the other hand, was designed to have the same long-term spectrum as that of the target speaker. Hence, energetic masking should be more effective with SSN than with TTB. These two effects appear to overwhelm the distracting effect produced by the semantic content of the babble. It was beyond the scope of this study to investigate these potential effects of gap listening and energetic versus informational masking. Future studies may be conducted to elucidate the roles played by these effects in CI users with SSD.

We also found a significant influence of masker on binaural effect comparing SRT with CI at 6 months to preoperatively unaided. Across spatial configurations, binaural benefit was 1.8 dB larger in TTB compared to SSN. Consistent with the significantly larger binaural benefit observed with CI in TTB compared to SSN, Bernstein et al. [29] found significant benefits with CI for speech in a same-gender TTB, but not for SSN in a squelch configuration presenting only the masker to the CI ear. Similarly, Dirks et al. [40] showed a larger benefit with CI for speech in time-reversed same-gender TTB compared to SSN, when assessing binaural effects post-operatively in an SSD cohort with at least 6 months device experience.

Longitudinal Effects of CI Use

Significant improvements in SRT with CI compared to preoperatively unaided, i.e., significant binaural effects were observed at different intervals in the different spatial configurations. Head shadow and summation benefits appeared at the earliest interval of 1 month in both maskers and were also present at 3 and 6 months. While there was no further change in summation effect from 1 to 6 months in either masker, head shadow effect increased in S0N±90 in TTB and in SSSDNAH in SSN from 3 to 6 months. Squelch benefit took longest to develop and was found at 6 months in S0NSSD in TTB only.

Galvin et al. [17] evaluated speech understanding in SSN in SSD CI recipients preoperatively unaided, and at 1, 3, and 6 months post-activation. Similar to our findings, significant longitudinal improvements were seen in the summation and the S0N±90° head shadow configuration. The head shadow benefits were evident by 3 months post-activation, while the summation benefits were evident by the first month, as in our study. No significant squelch effect was observed at any test interval, consistent with our findings in the SSN masker.

Buss et al. [28] assessed binaural effects in SSD CI recipients for speech recognition in 10-talker babble at intervals including preoperatively unaided, and 1, 3, 6, 9, and 12 months with CI. Relative to unaided preoperative performance, improvements in the S0N±90° head shadow and summation configurations were evident by 1 month and were significant by 12 months post-activation. Contrastingly, we found a significant squelch benefit in TTB with CI at 6 months, whereas they found no significant squelch effect at any test interval.

Regarding the emergence of binaural effects with CI over time, we found head shadow and summation benefits as early as 1 to 3 months, which was similar to previous studies [17, 28]. While no squelch effect appeared in the latter two studies, we observed a squelch benefit with CI, at least in TTB, at a later interval of 6 months. Our results are in line with previous findings in bilateral CI recipients who observed that squelch benefits emerged only after 1 year or more of CI experience [41]. Taken together, this suggests that squelch may have the most delayed period of development. A continued investigation with the same cohort from the present study will assess the extent to which binaural effects, and in particular squelch, continue to improve after the 6-month interval. As for the binaural effects, the large majority of the statistically significant differences in SRT over time was greater than 1 dB and thus clinically relevant.

Limitations

A limitation of the present study is the relatively small sample size due the prospective study design. The number of subjects with SSD seeking treatment within the scheduled recruitment period and fulfilling the study inclusion criteria was limited. To further support our findings, additional research and possibly pooling of data across multiple centers using a unified testing framework should be considered in the light of the consensus initiative by Van de Heyning et al. [5].

Following the unified testing framework for assessing hearing devices and SSD, this prospective study demonstrated that a CI – compared to CROS and BCD trial devices – provides users with better speech understanding in noise. With the CI, early significant improvements in speech understanding in noise were observed at both 1 and 3 months for head shadow and summation, while squelch benefits, if present, appeared later at 6 months post-activation. Further significant longitudinal improvements were observed between intervals in some target-masker spatial configurations. We found that short-term usage of both CROS and BCD yielded benefits in some configurations, but generally of a lower magnitude compared to CI at 6 months. For the squelch effect, CROS use led to a significant deterioration of speech understanding in noise in both spatial configurations and maskers. Lastly, we found that SSN was a more challenging masker compared to TTB for listening with CI.

The Freiburg authors thank the Fördergesellschaft “Taube Kinder lernen Hören e.V.” for its considerable support of the CI rehabilitation center in Freiburg in the past several years.

The study protocol was reviewed and approved by the Ethics Committee of the University of Freiburg, Germany (approval number 230/16). All procedures were in accordance with the World Medical Association Declaration of Helsinki (2013). Written informed consent was obtained from participants prior to enrollment.

T.W. reports grants and nonfinancial support from MED-EL Deutschland GmbH, during the conduct of the study; grants and nonfinancial support from Advanced Bionics AG, grants and nonfinancial support from Cochlear Deutschland GmbH & Co. KG, grants and nonfinancial support from MED-EL Deutschland GmbH, outside the submitted work. I.K. reports nonfinancial support from MED-EL Deutschland GmbH, during the conduct of the study. R.S. is an employee of MED-EL GmbH. S.B. is an employee of MED-EL Deutschland GmbH. A.A. and S.A. report grants and nonfinancial support from Advanced Bionics AG, grants and nonfinancial support from Cochlear Deutschland GmbH & Co. KG, grants and nonfinancial support from MED-EL Deutschland GmbH, grants and nonfinancial support from Oticon Medical A/S, outside the submitted work. L.J. and K.W. have no conflicts of interest to declare.

This work was supported by a research grant from MED-EL Deutschland GmbH.

T.W.: conceptualization, methodology, formal analysis, writing – original draft, and writing – review and editing. I.K.: methodology, investigation, formal analysis, and writing – review and editing. L.J. and K.W.: methodology and investigation. R.S. and S.B.: writing – original draft and writing – review and editing. A.A.: writing – review and editing. S.A.: conceptualization and writing – review and editing.

All data generated or analyzed during this study are included in this article. Further inquiries can be forwarded to the corresponding author.

1.
Giolas
TG
,
Wark
DJ
.
Communication problems associated with unilateral hearing loss
.
J Speech Hear Disord
.
1967
;
32
(
4
):
336
43
. .
2.
Wie
OB
,
Pripp
AH
,
Tvete
O
.
Unilateral deafness in adults: effects on communication and social interaction
.
Ann Otol Rhinol Laryngol
.
2010
;
119
(
11
):
772
81
.
3.
Slattery
WH
,
Middlebrooks
JC
.
Monaural sound localization: acute versus chronic unilateral impairment
.
Hear Res
.
1994
;
75
(
1–2
):
38
46
. .
4.
Nelson
E
,
Reeder
RM
,
Holden
LK
,
Firszt
JB
.
Front- and rear-facing horizontal sound localization results in adults with unilateral hearing loss and normal hearing
.
Hear Res
.
2019
;
372
:
3
9
. .
5.
Van de Heyning
P
,
Távora-Vieira
D
,
Mertens
G
,
Van Rompaey
V
,
Rajan
GP
,
Müller
J
, et al
.
Towards a unified testing framework for single-sided deafness studies: a consensus paper
.
Audiol Neurootol
.
2016
;
21
(
6
):
391
8
. .
6.
Távora-Vieira
D
,
Rajan
GP
,
Van de Heyning
P
,
Mertens
G
.
Evaluating the long-term hearing outcomes of cochlear implant users with single-sided deafness
.
Otol Neurotol
.
2019
;
40
(
6
):
e575
80
. .
7.
Kurz
A
,
Rak
K
,
Hagen
R
,
Ehrmann-Müller
D
.
Evaluating the decision for cochlear implantation in individuals with single-sided deafness (SSD); implementing the SSD consensus protocol into clinical routine
.
Otol Neurotol
.
2020
;
41
(
6
):
727
35
. .
8.
Vermeire
K
,
Van de Heyning
P
.
Binaural hearing after cochlear implantation in subjects with unilateral sensorineural deafness and tinnitus
.
Audiol Neurootol
.
2009
;
14
(
3
):
163
71
. .
9.
Arndt
S
,
Laszig
R
,
Aschendorff
A
,
Hassepass
F
,
Beck
R
,
Wesarg
T
.
Cochlear implant treatment of patients with single-sided deafness or asymmetric hearing loss
.
HNO
.
2017
;
65
(
Suppl 2
):
98
108
. .
10.
Mertens
G
,
De Bodt
M
,
Van de Heyning
P
.
Evaluation of long-term cochlear implant use in subjects with acquired unilateral profound hearing loss: focus on binaural auditory outcomes
.
Ear Hear
.
2017
;
38
(
1
):
117
25
. .
11.
Jakob
TF
,
Speck
I
,
Rauch
AK
,
Hassepass
F
,
Ketterer
MC
,
Beck
R
, et al
.
Bone-anchored hearing system, contralateral routing of signals hearing aid or cochlear implant: what is best in single-sided deafness
.
Eur Arch Otorhinolaryngol
.
2022
;
279
(
1
):
149
58
. .
12.
Távora-Vieira
D
,
Marino
R
,
Acharya
A
,
Rajan
GP
.
The impact of cochlear implantation on speech understanding, subjective hearing performance, and tinnitus perception in patients with unilateral severe to profound hearing loss
.
Otol Neurotol
.
2015
;
36
(
3
):
430
6
. .
13.
Mertens
G
,
Desmet
J
,
De Bodt
M
,
Van de Heyning
P
.
Prospective case-controlled sound localisation study after cochlear implantation in adults with single-sided deafness and ipsilateral tinnitus
.
Clin Otolaryngol
.
2016
;
41
(
5
):
511
8
. .
14.
Van de Heyning
P
,
Vermeire
K
,
Diebl
M
,
Nopp
P
,
Anderson
I
,
De Ridder
D
.
Incapacitating unilateral tinnitus in single-sided deafness treated by cochlear implantation
.
Ann Otol Rhinol Laryngol
.
2008
;
117
(
9
):
645
52
. .
15.
Arndt
S
,
Aschendorff
A
,
Laszig
R
,
Beck
R
,
Schild
C
,
Kroeger
S
, et al
.
Comparison of pseudobinaural hearing to real binaural hearing rehabilitation after cochlear implantation in patients with unilateral deafness and tinnitus
.
Otol Neurotol
.
2011
;
32
(
1
):
39
47
. .
16.
Holder
JT
,
O’Connell
B
,
Hedley-Williams
A
,
Wanna
G
.
Cochlear implantation for single-sided deafness and tinnitus suppression
.
Am J Otolaryngol
.
2017
;
38
(
2
):
226
29
. .
17.
Galvin
JJ
3rd
,
Fu
QJ
,
Wilkinson
EP
,
Mills
D
,
Hagan
SC
,
Lupo
JE
, et al
.
Benefits of cochlear implantation for single-sided deafness: data from the House Clinic-University of Southern California-University of California, Los Angeles clinical trial
.
Ear Hear
.
2019
;
40
(
4
):
766
81
. .
18.
Assouly
KKS
,
van Heteren
JAA
,
Stokroos
RJ
,
Stegeman
I
,
Smit
AL
.
Cochlear implantation for patients with tinnitus: a systematic review
.
Prog Brain Res
.
2021
;
260
:
27
50
. .
19.
Jacob
R
,
Stelzig
Y
,
Nopp
P
,
Schleich
P
.
Audiologische Ergebnisse mit Cochlear implant bei einseitiger Taubheit [Audiological results with cochlear implants for single-sided deafness]
.
HNO
.
2011
;
59
(
5
):
453
‒60. .
20.
Härkönen
K
,
Kivekäs
I
,
Rautiainen
M
,
Kotti
V
,
Sivonen
V
,
Vasama
JP
.
Single-sided deafness: the effect of cochlear implantation on quality of life, quality of hearing, and working performance
.
ORL J Otorhinolaryngol Relat Spec
.
2015
;
77
(
6
):
339
45
. .
21.
Dillon
MT
,
Buss
E
,
Rooth
MA
,
King
ER
,
Deres
EJ
,
Buchman
CA
, et al
.
Effect of cochlear implantation on quality of life in adults with unilateral hearing loss
.
Audiol Neurootol
.
2017
;
22
(
4–5
):
259
71
. .
22.
Häußler
SM
,
Köpke
V
,
Knopke
S
,
Gräbel
S
,
Olze
H
.
Multifactorial positive influence of cochlear implantation on patients with single-sided deafness
.
Laryngoscope
.
2020
;
130
(
2
):
500
6
. .
23.
van Zon
A
,
Peters
JP
,
Stegeman
I
,
Smit
AL
,
Grolman
W
.
Cochlear implantation for patients with single-sided deafness or asymmetrical hearing loss: a systematic review of the evidence
.
Otol Neurotol
.
2015
;
36
(
2
):
209
19
. .
24.
Brand
T
,
Wagener
KC.
Eigenschaften
,
Leistungen und Grenzen von Matrixtests [Characteristics, advantages, and limits of matrix tests]
.
HNO
.
2017
;
65
(
3
):
182
‒8. .
25.
Peters
JPM
,
van Heteren
JAA
,
Wendrich
AW
,
van Zanten
GA
,
Grolman
W
,
Stokroos
RJ
, et al
.
Short-term outcomes of cochlear implantation for single-sided deafness compared to bone conduction devices and contralateral routing of sound hearing aids-Results of a Randomised controlled trial (CINGLE-trial)
.
PLoS One
.
2021
;
16
(
10
):
e0257447
. .
26.
Hochmair-Desoyer
I
,
Schulz
E
,
Moser
L
,
Schmidt
M
.
The HSM sentence test as a tool for evaluating the speech understanding in noise of cochlear implant users
.
Am J Otol
.
1997
;
18
(
6 Suppl
):
S83
.
27.
Deep
NL
,
Spitzer
ER
,
Shapiro
WH
,
Waltzman
SB
,
Roland
JT
Jr
,
Friedmann
DR
.
Cochlear implantation in adults with single-sided deafness: outcomes and device use
.
Otol Neurotol
.
2021
;
42
(
3
):
414
23
. .
28.
Buss
E
,
Dillon
MT
,
Rooth
MA
,
King
ER
,
Deres
EJ
,
Buchman
CA
, et al
.
Effects of cochlear implantation on binaural hearing in adults with unilateral hearing loss
.
Trends Hear
.
2018
;
22
:
2331216518771173
. .
29.
Bernstein
JG
,
Goupell
MJ
,
Schuchman
GI
,
Rivera
AL
,
Brungart
DS
.
Having two ears facilitates the perceptual separation of concurrent talkers for bilateral and single-sided deaf cochlear implantees
.
Ear Hear
.
2016
;
37
(
3
):
289
302
. .
30.
Müller
V
,
Lang-Roth
R
.
Speech recognition with informational and energetic maskers in patients with single-sided deafness after cochlear implantation
.
J Speech Lang Hear Res
.
2021
;
64
(
8
):
3343
56
. .
31.
Grossmann
W
,
Brill
S
,
Moeltner
A
,
Mlynski
R
,
Hagen
R
,
Radeloff
A
.
Cochlear implantation improves spatial release from masking and restores localization abilities in single-sided deaf patients
.
Otol Neurotol
.
2016
;
37
(
6
):
658
64
. .
32.
Gartrell
BC
,
Jones
HG
,
Kan
A
,
Buhr-Lawler
M
,
Gubbels
SP
,
Litovsky
RY
.
Investigating long-term effects of cochlear implantation in single-sided deafness: a best practice model for longitudinal assessment of spatial hearing abilities and tinnitus handicap
.
Otol Neurotol
.
2014
;
35
(
9
):
1525
32
. .
33.
Gründer
I
,
Seidl
RO
,
Ernst
A
,
Todt
I
.
Wertigkeit der BAHA-Testung für das postoperative Hörergebnis [Relative value of BAHA testing for the postoperative audiological outcome]
.
HNO
.
2008
;
56
(
10
):
1020
24
. .
34.
Stenfelt
S
.
Transcranial attenuation of bone-conducted sound when stimulation is at the mastoid and at the bone conduction hearing aid position
.
Otol Neurotol
.
2012
;
33
(
2
):
105
14
. .
35.
Dieudonné
B
,
Francart
T
.
Redundant information is sometimes more beneficial than spatial information to understand speech in noise
.
Ear Hear
.
2019
;
40
(
3
):
545
54
. .
36.
Huber
AM
,
Strauchmann
B
,
Caversaccio
MD
,
Wimmer
W
,
Linder
T
,
De Min
N
, et al
.
Multicenter results with an active transcutaneous bone conduction implant in patients with single-sided deafness
.
Otol Neurotol
.
2022
;
43
(
2
):
227
35
. .
37.
Stronks
HC
,
Apperloo
E
,
Koning
R
,
Briaire
JJ
,
Frijns
JHM
.
SoftVoice improves speech recognition and reduces listening effort in cochlear implant users
.
Ear Hear
.
2021
;
42
(
2
):
381
92
. .
38.
Prejban
DA
,
Hamzavi
JS
,
Arnoldner
C
,
Liepins
R
,
Honeder
C
,
Kaider
A
, et al
.
Single sided deaf cochlear implant users in the difficult listening situation: speech perception and subjective benefit
.
Otol Neurotol
.
2018
;
39
(
9
):
e803
9
. .
39.
Fastl
H
.
A background noise for speech audiometry
.
Audiol Acoust
.
1987
;
26
:
2
13
.
40.
Dirks
C
,
Nelson
PB
,
Sladen
DP
,
Oxenham
AJ
.
Mechanisms of localization and speech perception with colocated and spatially separated noise and speech maskers under single-sided deafness with a cochlear implant
.
Ear Hear
.
2019
;
40
(
6
):
1293
306
. .
41.
Eapen
RJ
,
Buss
E
,
Adunka
MC
,
Pillsbury
HC
3rd
,
Buchman
CA
.
Hearing-in-noise benefits after bilateral simultaneous cochlear implantation continue to improve 4 years after implantation
.
Otol Neurotol
.
2009
;
30
(
2
):
153
9
. .