Computer-based auditory skill building program for aural (re)habilitation

描述（由申请人提供）：根据MarkeTrak VII调查（Kochkin 2008），美国大约有3500万人患有某种程度的听力损失。其中，近1000万人拥有助听器，67，000人拥有人工耳蜗。听觉（再）障碍（AR）计划的目标是建立听力损失患者的听力技能，以便他们能够理解其环境中的语言。有了合适的听力设备和AR，听力损失的人可以开始辨别和识别周围语言中的语音。 在这个项目的第一阶段，我们开发了一个自适应计算机引导的AR软件程序，称为声音识别导师（SIT），专注于声音识别和辨别。有关SIT的视频演示，请访问http://www.biospeech.com/research/ar.html。SIT提供了一个自动化的声音目标序列进行训练，它选择的目标类型取决于用户的性能。它还选择是否以听觉或视听模式呈现目标。SIT使用所有单词目标的自然人类视听记录。它还使用简单的单词儿童友好的图像，使这个软件适合儿童和乐趣。对24名儿童和成人进行的软件评估结果表明，在10次会议的过程中，受试者显着提高了他们的声音识别技能。儿童从自动化课程序列中受益比成人多。成年人倾向于选择他们自己的声音目标进行训练。此外，我们发现，受试者在他们训练的特定声音目标上有所改善，但前测试中没有训练的问题目标在后测试中没有改善。这表明，改善并不是由于增加对语言的接触，而是通过训练特定的声音目标。 在第二阶段的建议中，我们将扩展SIT并将其发展为可销售的产品。该建议包括以下目标：（一）建立一个西班牙语版的SIT;（二）扩展SIT，在句子中嵌入单词，并带有可选的背景噪音和多个发言者;（三）扩展SIT的开放式培训模块;（四）扩展SIT，以创建DVD安装程序，与网络连接，并开发家庭登录用户界面;（v）评估专为家居使用而设的计划; 公共卫生相关性：在美国，大约有3500万人有一定程度的听力损失，其中近1000万人有助听器，67，000人有人工耳蜗植入物，他们需要密集的听觉（再）障碍（AR）训练，以充分实现其假体的潜在益处并获得可理解的语音。与专家面对面的AR培训是昂贵的，并不总是由保险覆盖，并不总是可行的，因为客户可能无法获得专家，特别是如果他们说不同的语言。拟议中的项目将扩大一个自适应计算机引导的软件程序（语音识别导师，SIT），专注于学习音素辨别和识别，并将提供西班牙语版本的软件，提供额外的合成语音开放式培训，提供额外的自适应句子任务，其中个人接触到嵌入在句子中的单词，背景噪音，和多个扬声器，并将由更大的用户组评估家庭使用。

DESCRIPTION (provided by applicant): According to the MarkeTrak VII survey (Kochkin 2008), there are roughly 35 million individuals in the US with some degree of hearing loss. Of these, almost 10 million have hearing aids and 67,000 have cochlear implants. The goal of auditory (re)habilitation (AR) programs is to build the listening skills of individuals with hearing loss, so that they can comprehend the language in their environment. With appropriate hearing devices and AR, individuals with hearing loss can begin to discriminate and identify the speech sounds in the spoken language around them. In Phase I of this project we developed an adaptive computer-guided AR software program called the Sound Identification Tutor (SIT) that focused on sound identification and discrimination. For a video demonstration showcasing the SIT please visit http://www.biospeech.com/research/ar.html. The SIT provides an automated sequence of sound targets to be trained and it chooses the type of target depending on the performance of the user. It also chooses whether to present the target in auditory or audiovisual mode. The SIT uses natural human audiovisual recordings of all the word targets. It also uses child-friendly images of simple words to make this software suitable and fun for children. The results of an evaluation of the software with 24 children and adults showed that over a course of 10 sessions subjects significantly improved their sound identification skills. Children benefitted more from the automated lesson sequence than adults. Adults tended to prefer to choose their own sound targets to train on. In addition, we showed that subjects improved on specific sound targets they trained on, but problematic targets from the pre-test that were not trained on did not improve in the post-test. This showcases the fact that improvements were not caused by increased exposure to speech but by training specific sound targets. In this Phase II proposal, we will extend the SIT and develop it into a marketable product. The proposal consists of the following aims: (i) To Build a Spanish Version of the SIT; (ii) To Extend the SIT with Words Embedded in Sentences with Optional Background Noise and Multiple Speakers; (iii) To Extend the Open- Ended Training Module of the SIT; (iv) To Expand the SIT to Create a DVD installer, Interface with the Web, and Develop an In-home Login UI; (v) To Evaluate the Program Specifically for In-home Use; (vi) To Create a Built-in Assessment Protocol to Produce User-specific Listening Intervention Plans PUBLIC HEALTH RELEVANCE: In the United States, there are roughly 35 million individuals with some degree of hearing loss and of these, almost 10 million have hearing aids and 67,000 have cochlear implants who need intensive auditory (re)habilitation (AR) training to fully realize the potential benefit of their prostheses and attain intelligible speech. Face-to-face AR training with a specialist is expensive, not always covered by insurance and not always feasible because clients may not have access to a specialist, esp. if they speak a different language. The proposed project will expand upon an adaptive computer-guided software program (the Speech Identification Tutor, SIT) that focuses on learning phoneme discrimination and identification and will provide a Spanish version of the software, provide additional synthetic voices for open-ended training, provide an additional adaptive sentence task where individuals are exposed to words embedded in sentences, with background noise, and multiple speakers and will be evaluated for in-home use by a larger group of users.