中文摘要：

为了改进以正弦信号合成的电子耳蜗听觉仿真模型在仿真电子耳蜗植入患者所感知的语音信息的缺陷,建立了增加电极位置信息的正弦信号合成的听觉仿真模型.通过大样本量的不同汉语语言材料的听觉仿真实验,其中包括按元音、辅音、数字、音调和单字词等5种汉语语言材料,研究了基于连续交叉采样算法的不同实验条件下电极位置对言语识别率的影响,得出了汉语元音和单字词对植入深度更加敏感的结论.为进一步将这一仿真模型应用到电子耳蜗研究上奠定了基础.

英文摘要：

Problems occur in acoustic simulation models based on synthesizing sinusoidal signals to test cochlear implant patients＇ ability to recognize speech information.An improved acoustic simulation model was proposed that synthesizes sinusoidal signals that are adapted to electrode sites.A large number of experiments were run using different Mandarin materials,including vowels,consonants,numbers,tones and monosyllabic words.All experiments were based on continuous interleaved sampler（CIS） algorithm with different electrode insertion depths.Results demonstrated that Mandarin vowels and monosyllabic words were comparatively sensitive to electrode insertion depth.It established an importation foundation for electrode site information in cochlear implant acoustic simulation models.