Basic and clinical investigation of speech processing mechanism for the improvement of cochlear implant

1)语言感知(1)内耳语言加工（耳蜗电图）：正常被试在元音和CV音节的开头以及元音的连续部分都有CAPs，而CV音节的元音开头没有CAPs。部分感音神经性听力损失患者缺乏这些cap。我们确认了内耳产生的时间信息对语言感知的重要性，并开发了基于扩音和灵活时间窗的数字助听器。(2)口语皮层加工机制（PET、MEG、MRI）： PET研究表明，正常受试者各脑半球的听觉关联区主要处理来自对侧耳的语言输入，语音激活了听觉关联区，而噪音仅在初级听觉区引起rCBF增加。语前失聪患者的听觉关联区在语音刺激下也没有激活，这表明在没有语音刺激的情况下，语言皮质网络没有发育。通过脑磁图分析，我们发现正常受试者的听觉皮层在语音刺激下比在噪声刺激下更活跃，从而证实了大脑皮层改变语音和噪声的机制。我们还发现，对侧半球的反应在潜伏期出现得比正常受试者输入声音的同侧半球更早。相反，在单侧耳聋患者中，没有观察到这种偏侧。这些结果显示单侧听力丧失后皮层网络的重组。我们发现，正常成人受试者在不监控自己声音的情况下发出熟悉的语音，但当自己的声音听起来与清晰的声音不同时，听觉语言反馈变得活跃。这些结果表明，大脑皮层存在完全获得性的语音输出机制，以及听觉语言反馈系统对发声和发音的纠正机制。在这项研究中，我们证实了辅助运动区对规划发音运动和小脑对调节发声的重要性。同时也证明了大脑皮层对语音产生的加工过程具有顺序并行的时间过程，证明了大脑功能的合理性。少

1) Perception of language(1) processing of language in inner ear (Electrocochleography) : In normal subjects, CAPs were observed at the head of vowel and CV syllable sounds as well as at the continuous part of vowels, while CAPs were absent at the onset of vowel in CV syllable. These CAPs were lack in some sensorineural hearing loss patients. We confirmed the importance of temporal information made in inner ear for perception of language, and developed digital hearing aid based on amplification with flexible time window.(2) cortical processing mechnism for spoken language (PET,MEG,MRI) : PET studies showed that the auditory association areas in each hemisphere mainly processed language input from the contrlateral ear in normal subjects, and that speech sounds activated the auditory association areas while noise induced rCBF increase only in the primary auditory area in cochlear implant users. In prelingual deaf patiente, the activation of the auditory association areas was absent even … More with speech sounds stimulation, which indicated that the cortical network for language was not developed without speech sound stimuli. Using MEG, we found that the auditory cortices were more activated by speech sound than by noise in normal subjects, and confirmed the cortical mechanism to alter speech and noise. We also found that the response in the contralateral hemisphere appeared earlier in latency than that in the ipsilateral one to the ear where sounds were input in normal subjects. On the contrary, in unilateral deaf patients, this laterality was not observed. These results showed the reorganization of cortical network after unilateral hearing loss.2) cortical mechanisms for speech productionWe found that normal adult subjects vocalized familiar speech without monitoring their own voices in the auditory association areas, but that when the own voice sounded different to the articulated one, the auditory verbal feedback became active. These results indicate that there exist fully acquired speech output mechanism in cortex as well as the mechanism for correction of vocalization and articulation using auditory verbal feedback system. In this study, we confirmed the importance of the supplementary motor area for programing the articulatory movement and the cerebellum for the adjustment of vocalization. Wa also proved the time course of cortical processing for speech production to be sequential and parallel, which demonstrated the reasonable function of brain. Less