ANALYSIS OF SPEECH-PERCEPTION BY PATIENTS WITH A 22-CHANNEL COCHLEAR IMPLANT

Cochlear implant patients with F0/F1/F2 speech processing strategy can recognize vowels with 82% accuracy under the cochlear implant alone c ondition. For monosyllable consonants, they can perceive with 50% accu racy under the cochlear implant alone condition, and with 85% accuracy under the conditions of cochlear implant plus lip-reading. Informatio n transfer ratios for monosyllable consonants ranged from 23.4% to 54. 5% for the cochlear implant alone condition. For the conditions of coc hlear implant plus lipreading, the range of ratios was from 64.2% to 8 7.3%. This indicates that the lip-reading plays an important role for the cochlear implantees. Multiple regression analysis revealed that na sality and semivowels contributed more than other consonant features t o correct answer rates for the cochlear implant alone condition. The p lace of articulation contributes most to correct answer rates under th e conditions of cochlear implant plus lip-reading. These studies demon strate that visual information supplements the deficit in auditory inf ormation for cochlear implantees. Test results for monosyllable, word and sentence recognition revealed that scores of lip-reading decreased in order of monosyllable, word and sentence, whereas scores of cochle ar implant alone showed no definite change. From the correlation analy sis, it is concluded that the cochlear implant plays a main role for r ecognition of monosyllables, words and sentences, while lip-reading in creases its weight in order of monosyllables, words and sentences. Thi s indicates that visual information is effective for the perception of syntax by the brain, and that it complements auditory information. Fr om the comparative study of the formant extraction strategy and the en velope tracking strategy, it was definitely shown that the latter prov ides better speech sound perception than the former, especially in noi sy conditions. Because the acoustic nerve activity represents an envel ope of speech sounds, the envelope tracking strategy is closer to the function of the cochlea than formant extraction strategy. Therefore, t he brain centers for speech recognition can adjust more easily to this pattern of activation. I believe that the new type of cochlear implan t is an important innovation in the cochlear implant, and that it will progress into an even more effective tool for deaf people.