| 摘要: | 隨著環境噪音的增加與人口老化趨勢，有聽力損傷的人數不斷增加，根據內政統計目前台灣因聽覺機能損傷者領有身心障礙手冊者約十萬多人，對於重度或極重度 ( 90dB 以上者 ) 聽損者，以目前的技術，人工電子耳是恢復重度聽損者聽覺另一種有效的方法。其中，人工電子耳的效能受語音編碼策略影響最大。目前人工電子耳對於噪音環境下語音辨識及音樂辨識的效能，即使選擇最適當的語音編碼策略下仍有待改進。加上中文語音的辨識偏重於語音基本頻率與週期性訊息的感知。因此，本研究目的是想瞭解若以助聽器提供低頻信號合併人工電子耳( EAS, combined electrical and acoustic stimulation)使用下，在噪音環境中如何對中文字詞及聲調的辨識、音高鑑別及旋律辨識的產生影響。在第一年的計畫中，我們模擬不同刺激速率下對五位正常聽力者(4男1女) 進行中文聲調和單字辨識率的測試。測試結果用配對雙尾t檢定發現（paired two-tailed t-test），250Hz、900Hz、1200Hz及1800Hz在聲調及單字辨識率上沒有顯著差異。同時在不同刺激速率各聲調辨識率的平均值中，以一(56.8%)、二聲 (57%) 效果最差，三聲 (90.3%) 其次，而四聲(93.6%)效果最好（董書豪, 民96）。這結果和臨床文獻 (許筱曼, 民91) 比較是相同的。另外雙耳聽在8位正常聽力者上單字及聲調辨識率皆獲得提升（單字辨識率平均值由32％上升至63％，聲調辨識率平均值由60％上升至98％）。其後再將此一結果和4位實際使用進階結合編碼策略之人工電子耳配戴者做一比較，其平均值結果在單字及聲調辨識率上有提升（單字辨識率平均值由5％上升至22％，聲調辨識率平均值由55％上升至73％）。最後結果也發現，若能提供頻域成分愈多的原始訊息，在單字辨識率上會得到更多助益。本計劃將繼續使用人工電子耳 ACE (Advanced combined encoder) 語音編碼策略模擬平台，將個案數增加到10名正常聽力 (頻率範圍在250Hz~8000Hz間的聽力小於 20dBHL)、使用國語的大專生及7-10名配戴人工電子耳及助聽器的聽損者為受測對象。對於正常聽力者給予電腦模擬人工電子耳及合併助聽器的語音信號與歌曲作測試，並以聽損個案作臨床實驗對照，研究人工電子耳在頻域(spectral cues)及時域(temporal cues) 的相關參數對中文語音辨識、音高鑑別及旋律辨識的影響，及EAS對於噪音環境中對中文語音辨識及音樂辨識的改善率。 ; As aging of population and environmental background noise in our community increases, population with various degrees of hearing loss also increases. Based on the demographic data of Ministry of Interior, R.O.C., there are more than one hundred thousand people diagnosed as hearing disabled with various degrees of hearing loss. For people with profound hearing loss (greater than 90dBHL), cochlear implants (CI) could be an effective approach for them to hear the speech again. Among various factors that will affect the performance of cochlear implant, selection of speech encoding strategy might affect the efficacy of cochlear implants. Even with appropriate speech encoding strategy, efficacy of speech recognition and music perception with CI under environmental background noise conditions are still the problems to be improved. In addition, recent studies have shown that fundamental frequency and periodic information of speech signal are important for Mandarin speech recognition and perception. Therefore the main purpose of this study is to understand how the combined electrical and acoustic stimulation (EAS) will affect Mandarin word and tone recognition, pitch identification, and melody recognition under the environmental background noise for patients with CI. For the first year project, five subjects (4 males and 1 female) with normal hearing were tested with simulated signals of four different stimulus rates (250Hz、900Hz、1200Hz 1800Hz). Preliminary results show that there is no statistically significant difference (paired two-tailed t-test) among these four stimulus rates for Chinese tones and words recognition rates. These results are consistent with previous clinical studies from other research. In addition, our mean recognition scores of Tones 1 (56.8%) and 2 (57%) are the lowest among four Chinese tones. Followed by the Tone 3 (90.3%), and the score of Tone 4 (93.6%) is the highest. In addition, we showed that the Chinese word and tone recognition rates of the NH listeners and CI recipients are statistically significantly improved (word recognition rates were improved from 32% to 63%, 60% to 98% for NH listeners and 5% to 22%, 55% to 73% for CI recipients). These results suggest that Chinese speech recognition could be enhanced if more speech information are provided from the HA to CI users and recognition rates are further improved with increasing amounts of frequency acoustic information. This study will develop a simulation platform for ACE (Advanced combined encoder) strategy of CI. Mandarin Monosyllabic Lexical Neighborhood Test (LNT) developed by the Department of Otolaryngology, National Cheng Kung University and ten well-known children songs will be used for test material. The simulated EAS speech signal and music under the environmental background noise will be tested on 10 Mandarin-speaking college students with normal hearing (250-8000 Hz, less than 20dBHL). For comparison purpose, Mandarin word and tone recognition and music perception under the environmental background noise will also be tested on 7-10 patients with CI. Results of this study will be useful for understanding how the parameters of spectral and temporal cues will affect word and tone recognition, pitch identification, and melody recognition and how EAS will improve Mandarin speech recognition and music perception under environmental background noise. ; 研究期間 9708 ~ 9807 |
