SPOKEN LANGUAGE PROCESSING WITHOUT AUDITION

一些严重听力受损(HI)的演讲者(唇读器)实现了 连接语音的准确率级别大于80%的单词 对，是这样。这些水平的性能是值得注意的，因为 现有的演讲阅读文学以视觉语音为特征 刺激措施非常模棱两可，演讲稿阅读器的词典大多 充满同义的(即，视觉上相似，但声音不同) 字里行间。这项研究计划的长期目标是提供一个 对知觉、语言和认知过程的综合描述 潜在的视觉言语知觉和表征相似之处 以及不同类型个体在这些过程中的差异 听觉和语言经验。为了启动这项工作，我们建议 将研究1)视觉语音的四个相互关联的项目 刺激，2)词的同源效应，3)词的效应 知识和词汇表征；4)连词的作用 关于单词识别的语音上下文。将进行直接比较 在四类受试者的高度熟练的讲稿阅读器中， 患有：听力正常、先天性听力损失、早发性听力 失聪和成人起病的听力损失。将采用的方法学 拟议的研究计划将从听觉领域改编 语音感知和口语单词识别。在某种程度上， 演讲稿阅读的结果复制了现有的 文学，可以得出结论，朗读演讲稿只是另一回事 人类一般的语言感知能力的表现， 通常完全归因于试镜。获得的知识作为 拟议研究的结果将应用于 HI儿童的言语和语言训练、朗读和双旋转 爱滋病。这些成果将为以下工作提供有价值的规范性数据库 各自的研究对象群体，并直接申请 构建高效的演讲阅读测试。我们对人类社会的研究 被试对词汇的知觉表征和词汇表征也会 对阅读有影响，对个人成功的活动也有影响 早期听力损失与语音知识高度相关 可能是通过阅读演讲稿获得的。

Some profoundly hearing-impaired (HI) speechreaders (lipreaders) achieve levels of accuracy with connected speech that are greater than 80% words correct. These levels of performance are remarkable given that the existing speechreading literature characterizes the visual phonetic stimulus as highly ambiguous and the speechreader's lexicon as mostly filled with homophenous (i.e., visually similar but different sounding) words. The long term goal of this research program is to provide an integrated account of the perceptual, linguistic, and cognitive processes underlying visual speech perception and to characterize the similarities and differences in those processes among individuals with different types of auditory and linguistic experience. To initiate this work we propose four interrelated projects that will study 1) the visual phonetic stimulus, 2) the effects of word homophene, 3) the effects of word knowledge and lexical representation, and 4) the effects of connected speech contexts on word recognition. Direct comparisons will be made among highly proficient speechreader in four populations of subjects, those with: normal hearing, congenital hearing loss, early-onset hearing loss, and adult-onset hearing loss. Methodologies to be employed in the proposed research program will be adapted from the domains of auditory speech perception and spoken word recognition. To the extent that results form speechreading replicate similar ones in the existing literature, it may be concluded that speechreading is but another manifestation of a general human capacity for speech perception, frequently attributed exclusively to audition. Knowledge obtained as a consequence of the proposed research will have applications for the speech and language raining of HI children, speechreading and birotactile aids. Results will provide valuable normative databases for the respective subject populations and have direct application for construction of efficient speechreading tests. Our studies of the perceptual and lexical representations of words by HI subjects will also have implications for reading, and activity whose success in individuals with early hearing loss is highly related to phonological knowledge that may be acquired via speechreading.