Speech Perception Assessment and Training System for Aphasia Rehabilitation

描述（由申请人提供）：本项目将测试计算机语音感知评估和训练系统（SPATS-AR）作为一种新型失语症康复选择的可行性。SPATS-AR将通过修改现有的软件来开发，这些软件已被发现在提高听力受损听众（SPATS-HI）和成人英语学习者（SPATS-ESL）的语音识别方面取得了成功。将确定使该系统适应失语症康复所需的修改，并评估修改后的系统的潜力。该项目特别关注提高语音感知能力。这有望更广泛地提高语言理解能力，因为语言感知障碍似乎对更高水平的语言有级联效应，对单词和句子的理解产生负面影响。SPATS提供客观的、自动化的训练，用于识别英语口语的组成部分、音节成分，以及识别有意义的自然口语句子。识别音节组成部分（起音、音核和音尾）的培训旨在为客户提供必要的技能，帮助他们攻克困难的单词并监控自己的语言。这些声音的重要性是根据它们在词汇和文本中的出现频率来排序的。在培训中有三个特别的特点被证明是有效的：(a)根据项目的重要性和客户的表现逐步介绍项目；(b)自适应项目选择（AIS），一种选择训练项目的专有方法；(c)“回应后重听”选项，允许客户将目标声音与错误的声音进行比较。与成分训练交织的是一个独特的句子测试和训练系统，它提供了自上而下的认知（如工作记忆）和自下而上的语言技能的练习。重要的是，在失语症中，理解问题不仅与语言处理困难（亚词汇、词汇和/或句法水平）有关，还与伴随的认知缺陷（如注意力、工作记忆）有关。因此，SPATS似乎非常适合针对失语症中经常受损的听力、阅读和认知技能。在第一阶段，患有轻度至中度失语症的参与者将从印第安纳大学言语和听力诊所招募。可以预见的是，音节组成任务将需要通过使用更小的发音集来大大简化。句子任务需要通过添加具有更简单句法结构的新句子、增加信噪比、简化目标词和陪衬词的显示来进行修改。我们的策略是首先采用与ESL和HI人群类似的刺激和程序，然后根据失语症患者的表现改变程序。预计在第二阶段最后确定的SPATS-AR将是一种有价值的康复选择，不仅因为其内容，而且因为其在家实践的远程康复潜力。

DESCRIPTION (provided by applicant): This project will test the feasibility of a computerized speech perception assessment and training system (SPATS-AR) as a novel aphasia rehabilitation option. SPATS-AR will be developed by modifying existing software that has been found to be successful in improving speech recognition by hearing-impaired listeners (SPATS-HI) and by adult learners of English (SPATS-ESL). The modifications needed to adapt the system for aphasia rehabilitation will be identified and the potential of the modified system will be evaluated. The project has a special focus on improving speech-perception capacities. This has the promise of improving language comprehension skills more broadly as speech-perception impairments appear to have a cascading effect on higher levels of language, negatively impacting comprehension of words and sentences. SPATS provides objective, automated training in the identification of the building blocks of spoken English, syllable constituents, and in the recognition of naturally spoken meaningful, sentences. Training in identifying the constituents of syllables (onsets, nuclei, and codas) is designed to give clients necessary skills for attacking difficult words and monitoring their own speech. These sounds have been ranked in their importance based on their lexical and textual frequencies of occurrence. Three special features have proven effective in training: (a) the graduated introduction of items based the item's importance and the client's performance; (b) Adaptive Item Selection (AIS), a proprietary method of selecting items for training; and, (c) the "post-response rehearing" option that allows the client to compare the target sound with the one mistaken for it. Interleaved with constituent training is a unique sentence testing and training system, which gives practice in top-down cognitive (e.g., working memory) and bottom-up linguistic skills. Importantly, in aphasia, comprehension problems have been found to be related to not only linguistic processing difficulties (at sublexical, lexical, and/or syntactic levels), but also to concomitant cognitive deficits (e.g., attention, working memory). Thus, SPATS appears well suited to target the listening, reading, and cognitive skills that are frequently compromised in aphasia. In Phase I, participants with mild to moderate aphasia will be recruited from the Indiana University Speech and Hearing Clinic. It is anticipated that the syllable constituent tasks will need to be greatly simplified by the use of smaller sets of sounds. The sentence task will need to be modified by adding new sentences with simpler syntactic constructions, increasing the signal-to-noise ratios, and simplifying the display of target and foil words. The strategy will be to begin with stimuli and procedures similar to those now used with ESL and HI populations and then alter the procedures as dictated by the performance of aphasic clients. It is anticipated that SPATS-AR, as finalized in Phase II, will be a valuable rehabilitative option not only because of its content but also because of its telerehabilitation potential for at-home practice. PUBLIC HEALTH RELEVANCE: This project will test the feasibility of a computerized Speech Perception Assessment and Training System (SPATS) as a novel aphasia treatment protocol. SPATS has proven effective in training both hearing- impaired listeners and learners of English as a second language to improve their ability both to identify English speech sounds and to recognize naturally spoken, meaningful sentences. A version of SPATS specifically designed to meet the needs of individuals with aphasia has the potential to improve these individuals' word recognition, comprehension, and cognitive skills and to offer a telerehabilitation option that may increase the efficiency of therapy, foster at-home practice, and bring treatment activities to those unable to regularly attend speech-language therapy sessions.