Phoneme recognition by a non-native English-speaking cochlear implantee using the WSP, MSP, and spectra speech processors: A case study

Abstract

This study presents the performance of a non-native English-speaking adult using the Nucleus 22 Channel cochlear implant. Included are the subject’s (1) history, (2) promontory stimulation results, (3) audiological information, (4) pre- and post-implantation speech recognition test results, (5) post-implantation auditory phoneme recognition using the wearable speech processor (WSP), the minispeech processor (MSP), and the Spectra processor, and (6) the subject’s subjective reports of benefits in terms of communication and social interactions. The subject, a highly motivated and intelligent male who, presented with a profound bilateral, progressive sensorineural hearing loss of unknown etiology, was 59 years old when the right ear was implanted. Prior to implantation the subject was aided in the left ear, but had never been aided in the right ear. The subject’s performance using each of three processors is presented for medial vowel recognition and medial consonant recognition using auditory stimulation only. Results are displayed through confusion matrices generated based on the subject’s responses to these acoustic stimuli. In summary, the vowel and consonant phoneme recognition of a subject with the Nucleus cochlear implant using various processors is presented. Improved phoneme recognition is demonstrated with filter banking versus feature extraction coding strategies.