Algorithms to improve listening in noise for cochlear implant users

In this study, a two-microphone noise reduction algorithm that aims to improve directional beamformer performance for cochlear implant (CI) users is presented. The algorithm is computationally inexpensive and estimates a spatially-based signal-to-noise ratio (SNR), attenuating time-frequency elements that have poor SNR. The attenuation function was specifically tuned for application to CI. Using a real-time implementation of the algorithm, evaluation took place in a variety of noisy situations in which the competing speech sources were spatially separated and changed location during the test. Speech intelligibility tests with CI users revealed the new algorithm improved speech performance by 4.6 dB in speech reception threshold (SRT) compared to a commercially available adaptive beamformer.