Word Recognition for Low-and High-pass Filtered Korean CVC
Monosyllables in Cochlear Implant Users

Jang, Hyunsook; Lee, Jiyeon; Yoo, Sooyeon; Jung, Eunjo; Park, Hongjoon

doi:2007.3.1.58

Audiol > Volume 3(1); 2007 > Article

Research Paper

Audiology 2007;3(1): 58-63.
Published online: June 30, 2007.
doi: http://doi.org/10.21848/audiol.2007.3.1.58

인공와우 착용자의 주파수 여과에 따른 일음절 재인 능력

장현숙¹, 이지연^1,2, 유수연^1,3, 정은조¹, 박홍준²

¹한림대학교 자연과학대학 언어청각학부, 청각언어연구소
²소리이비인후과
³경희대학교 이비인후과학교실

Word Recognition for Low-and High-pass Filtered Korean CVC Monosyllables in Cochlear Implant Users

Hyunsook Jang¹, Jiyeon Lee^1,2, Sooyeon Yoo^1,3, Eunjo Jung¹, Hongjoon Park²

¹Division of Speech Pathology and Audiology, College of Natural Science, Hallym University, Chuncheon, Korea
²Soree Ear Clinic, Seoul, Korea
³Department of ORL-HNS, College of Medicine, Kyunghee University, Seoul, Korea

Correspondence

Hyunsook Jang ,Tel: (033)248-2218, Fax: (02)542-4207, Email: hsjang@hallym.ac.kr

Received: April 16, 2007; Accepted: May 27, 2007. Published online: June 30, 2007.

ABSTRACT

Although spectral information is reduced or degraded, most normal listeners easily recognize spoken language because of the spectral redundancy of speech information. Little is known about speech intelligibility in cochlear implant listeners when the spectral resolution of speech signal is significant limited or the fine spectral structures are not available. This study examined the ability to recognize filtered speech in normal and cochlear implant listeners. Two experiments were conducted. In Experiment 1, Korean CVC monosyllabic words spoken by a female speaker were low- and high-pass filtered and presented to 20 normal-hearing subjects. Word recognition performance was obtained at most comfortable level for four low-pass filtered conditions (800-, 1200-, 1500-, and 1700-Hz cutoffs) and for four high-pass filtered conditions (1700-, 2100-, 2500-, and 3000-Hz cutoffs). In Experiment 2, two CVC monosyllabic lists that showed 90-100% correct recognition at low- and high-pass filtered conditions in normal-hearing subjects were selected and presented to individuals that had post-lingual deafness. Word recognition scores were measured in unfiltered conditions and in lowand high-pass filtered conditions. Ten individuals that had post-lingual deafness and used a cochlear implant participated in this study. Recognition performance was significantly poorer in CI listeners than in normal-hearing subjects. For the CI listeners, the overall percentage of correct recognition scores in unfiltered conditions was about 50%. In low-pass filtered conditions, recognition performance deteriorated with decreasing filter cutoffs. In high-pass filtered conditions, performance was extremely poor even in the 1700-Hz cutoff condition. Reduced spectral information cause significant difficulties in speech recognition in normal-hearing subjects and further worsened in CI listeners. This finding suggests that high degrees of spectral redundancy may not be available for cochlear implant users. CI users may equally have dependence on both low- and high-frequency hearing channels for speech recognition.

Key Words: Cochlear Implant, Filtering, Low-pass filter, High-pass filter

중심단어: 인공와우, 필터, 저주파수필터, 고주파수필터