Background: Cochlear implant (CI) users have difficulty understanding speech in background noise. Directional microphones, also known as beamformers, are spatial filters that enhance speech intelligibility in noisy environments by attenuating sounds originating from the sides and rear. Beamformers can be monaural or binaural. Monaural beamformers generate a directional beam on only the cochlear implant. For binaural beamforming, a second device, such as a contralateral routing of signals (CROS) device, can be incorporated, by capturing sounds presented to the ear contralateral to CI ear (CROS-ear) and rerouting them to the CI, which subsequently enhances speech intelligibility in noise. Multiple research groups have investigated these beamformers for CIs with varying results. However, the variability in the test setups used complicates the comparison of different studies. The extent to which these different noise configurations influence the effectiveness of directional microphones is not known.

Rationale: Assess the influence of noise configuration on the effectiveness of beamformers in CI users.

Methods: 18 postlingually deaf CI users, monaurally implanted with an Advanced Bionics device, participated in this study. Speech recognition thresholds (SRTs) were assessed using the Dutch/Flemish Matrix test. The SRT was defined as the signal-to-noise ratio (SNR) where 50% of speech was recognized correctly. Four conditions were tested for CI users: CI-only, CI-CROS, CI with UltraZoom (monaural beamformer), and CI-CROS with StereoZoom (binaural beamformer). Four different noise setups were tested: a homogeneous noise field, 8-loudspeaker ring, 3-loudspeaker ring, and 1-loudspeaker setup. Speech was always presented from a loudspeaker positioned directly in front of the participant. Noise level was 60 dBA on average. The CROS-ear was plugged to minimize influence of residual hearing.

Results: No significant differences were observed between the various noise setups in the CI-only and CI-CROS condition. UltraZoom and StereoZoom significantly improved speech understanding compared to the CI and CI-CROS condition (all comparisons p < 0.001). Specifically, improvements were 5.5 and 5.8 dB SNR in 3-loudspeaker, 3.3 and 4.2 dB SNR in 8-loudspeaker, and 3.8 and 4.5 dB SNR in the homogeneous noise field, respectively. Activating UltraZoom resulted in significant differences in SRTs across all noise setups. Similarly, with StereoZoom enabled, SRTs differed significantly across noise conditions, with the exception of the comparison between the 8-loudspeaker and homogeneous noise fields.

Conclusion: In all noise configurations, activating a beamformer led to increased speech intelligibility in CI users. However, the type of noise setup and the direction from which noise originates used during speech-in-noise testing does influence speech understanding, which must be considered when comparing studies.