Ten years of research on the sound quality of cochlear implants.

This report summarizes data collected over the span of more than a decade from four laboratories on the sound quality of cochlear implants. Sixteen projects (ten published and six unpublished) are summarized; two video and eleven audio files accompany the summaries. Fourteen of the projects used single sided deaf patients (SSD) fit with an implant (CI) as listeners; one project used SSD-CI patients with preserved hearing (HP) in their implanted ear; one project used conventional CI patients. Approximately a third of the SSD-CI patients could create close matches to CI sound quality. The matches varied as a function of electrode length. Patients fit with shorter electrode arrays used upshifted formant frequencies and/or voice pitch and spectral smearing to match sound quality. Approximately half of the patients with longer electrode arrays used low- or band-pass filtering, spectral smearing and/or pitch contour flattening to match sound quality. Other patients chose very few modifications - these signals were relatively clean. For the SSD-CI/HP patients, stimulation directed to both the CI and acoustic hearing was characterized, relative to stimulation directed only to the CI, by one or more of the following: (i) a lower F0, (ii) a less flattened F0 contour, (iii) less upshift in formant frequencies, (iv) less smear, and (v) fewer instances of high-pass filtering. The final project established that muffled sound quality could be reduced by including an input gain for mid frequencies.