Cross-frequency integration for speech recognition in bimodal hearing

描述（由申请人提供）：在许多研究中已经证明了电和声联合刺激在语音和音高感知方面的好处。在不同的研究、不同的测试材料和条件下，对报告的综合获益量的仔细检查是不同的。一些患者没有表现出任何综合效益，在极少数情况下，一些患者甚至表现出两种刺激之间潜在的不相容。已经有几次尝试将声学和电相结合的听力益处与残余听力中某些听觉功能的测量联系起来，但没有发现显著的相关性。本研究的重点是双峰听力，听者接受电刺激在一只耳朵和声刺激在对侧耳朵。本项目的长期目标是：(1)了解接受声电联合刺激的CI听者的语音处理过程；(2)为制定改善CI听者语音识别的康复策略提供依据。具体目的是：(1)识别高频区域电听力提取的语音信息和低频区域残留的声学听力；(2)研究从每只耳朵提取的信息如何在正常听力和人工耳蜗听者中整合；(3)将音素识别性能与句子识别性能联系起来。我们将应用几个成熟的语音集成模型，包括简单概率模型、感知模糊逻辑模型、预标记和后标记模型来预测综合听力表现的可理解性分数。这种基于模型的方法提供了一种方法，可以系统地研究人工耳蜗听者同时从声学和电刺激中提取语音信息并将这些信息整合到耳朵上的能力差异。这项工作具有很高的临床意义，因为它可以帮助识别植入物使用者在个人基础上遇到的信息提取和/或整合方面的缺陷，并有助于制定针对个人需求的康复策略。

DESCRIPTION (provided by applicant): The benefit of combined electric and acoustic stimulation in speech and pitch perception has been demonstrated in a number of studies. A close examination of the amount of combined benefit reported varies across studies and across test materials and conditions. Some patients did not demonstrate any combined benefit and in rare occasions, some even exhibited a potential incompatibility between the two types of stimulation. Several attempts have been made to relate combined acoustic and electric hearing benefits to some measure of auditory function in the residual hearing, but significant correlations were not found. This study focuses on bimodal hearing in which listeners receive electric stimulation in one ear and acoustic stimulation in the contralateral ear. The long-term goals of this project are (1) to understand the processing of speech in CI listeners who receive combined acoustic and electric stimulation, and (2) to provide a basis for the development of rehabilitation strategies for improving speech recognition in CI listeners. The specific aims are (1) to identify the speech information extracted in electric hearing in the high-frequency regions and residual acoustic hearing in the low-frequency regions; (2) to investigate how the extracted information from each ear is integrated in normal-hearing and cochlear-implant listeners; and (3) to relate phoneme recognition performance to sentence recognition performance. We will apply several well-developed speech integration models, including a simple probabilistic model, Fuzzy Logic Model of Perception, and Pre-Labeling and Post- Labeling models to predict intelligibility scores for combined hearing performance. This model-based approach provides the means to systematically study the differences in the abilities of cochlear-implant listeners to simultaneously extract speech information from acoustic and electric stimulation and integrate this information across ears. The proposed work is of high clinical relevance because it may help identify deficits in information extraction and/or integration, encountered by implant users on an individual basis and aid in developing rehabilitative strategies tailored to individual needs. Relevance: The purpose of this study is to investigate how speech information is integrated across ears in individuals who wear a cochlear implant in one ear and a hearing aid in the opposite ear. The proposed work is of high clinical relevance because it may help identify the problems encountered by implant users and aid in developing rehabilitative strategies tailored to individual needs.