人工電子耳為數十萬無法有效使用助聽器的聽損者帶來了重大的改變，然而，現今的電子耳聲音品質對於聲調語言、噪音環境中的語音以及音樂等方面仍有其限制。本計畫將整合三個以聽覺生理為基礎的電子耳聲音處理方法至本實驗室現有的人工電子耳實驗平台，並探討這些方法以及進階聯合編碼(ACE, advanced combination encoder)策略在不同組合的情況下對於中文語音辨識的效果。這些針對電子耳使用者缺乏的聽覺處理機制發所展出的方法，包括了參考基底膜的瞬間壓縮以及內側耳蝸橄欖反射(medial olivocochlear reflex)的生理助聽器演算法BioAid、依照聽覺神經纖維突觸適應效應的包絡線增強法EE (envelope enhancement) ，以及保留聽覺所需之週期性音高的基本頻率調變法F0mod (fundamental frequency modulation)。本計畫在第一年將於實驗平台上完成聲音處理方法的實作，並在多重的方法組合與噪音條件下進行模擬實驗。而在第二年的臨床實驗中，將擴充實驗平台的硬體控制功能，將處理過的信號將透過NIC (Nucleus implant communicator)工具與研究用處理器(research processor)直接串流至電子耳使用者體內的植入裝置，以評估實際的電刺激效果。透過模擬和臨床實驗結果，本計畫將比較各種處理方法的組合對於中文聲調和在噪音環境中語音辨識的影響，以期有助於電子耳聲音處理相關領域的研究發展。 ;Cochlear implants (CI) have significantly changed the lives of hundreds of thousands hearing impaired people who could not use hearing aids effectively. However, the sound quality of the cochlear implant today is still limited in the performance of tonal languages, speech in noisy conditions and music perception. This project is going to integrate three auditory physiology based sound processing methods into an existing CI simulation platform developed by our research team and investigate the Mandarin speech recognition performance of these methods with the ACE (advanced combination encoder) strategy under various combinations. The sound processing methods, developed from auditory processing mechanisms absent in CI users, involves a biologically inspired hearing aid (BioAid) algorithm models the instantaneous compression of the basilar membrane, the envelope enhancement (EE) strategy simulates the olivocochlear (MOC) reflex, and the F0 modulation (F0mod) strategy, which preserves the periodicity pitch required by auditory perception. In the first year of this project, sound processing methods are going to be implemented to the experiment platform, which will be used for CI stimulations with multiple combinations of processing methods and noisy conditions. In clinical experiments of the second year, the experiment platform will be expanded with hardware control functionality, which directly streams processed signals into a CI recipient’s internal implant via the NIC (Nucleus implant communicator) and a research processor. Hence the actual electric stimulation results can be evaluated. Based on the outcomes of simulations and clinical experiments, Mandarin lexical tone identification and speech recognition in noise affected by different combinations of sound processing methods are going to be compared in order to bring positive effects on related studies in the field of cochlear implant sound processing.
