The role of stimulus properties on past tense acquisition in children with SLI

描述（由申请人提供）：特殊语言障碍儿童（SLI）是一种没有任何明显原因的口语障碍，约占人口的7%，与同龄人相比，他们的长期学业和社交成绩明显较差。在学龄前，语法形态和句法的习得和使用方面的缺陷是最常见的缺陷之一。目前对词法和句法的干预效果有限（Law et al., 2004）。增强的语言干预技术将改善这一人群的长期积极结果。这项研究资助利用典型语言学习的输入驱动方法，该方法提出，通过提供稳定的范例，促进将语法结构的意义与形式联系起来，高频率的例子有助于促进语法结构的习得。同样，单词学习理论和人工语法学习理论表明，与高度典型的范例相比，使用非典型范例可以在更大程度上促进学习。该研究项目将这些理论扩展到干预中，通过使用计算建模和行为方法来研究刺激选择和表现对干预结果的作用。将进行两项早期疗效研究。其中一个重点是教有特殊语言障碍的儿童使用过去时，使用两种不同的方式选择动词用于干预。第二个研究重点是使用三种不同的输入分布，向特殊语言障碍儿童和普通语言儿童教授一种新的语素。行为研究的结果将与计算模型相联系，使我们能够测试输入结构和选择刺激的方法，这些方法超出了行为检查的可行性。这将使我们有可能测试关于什么能比单独使用传统行为方法更有效地改善干预措施的假设。研究者有丰富的经验与儿童的特殊语言障碍和评估使用语法形态的儿童。然而，她在计算建模、干预研究设计以及向语言病理学家传达结果方面缺乏专业知识。这项拨款的培训部分将确保研究者能够使用与行为结果相一致的计算建模技术，以更好地了解儿童的学习方式。它还将提高研究者独立开展干预研究的能力，并将研究结果传达给从业者。爱荷华大学以在特殊语言障碍和计算建模领域的工作而闻名，并且在临床相关研究方面有着悠久的历史。心理学和沟通障碍研究人员之间的现有合作使这里成为进行这种培训的理想环境。培训机会将帮助研究者开展独立的研究事业，研究改善语言障碍治疗的方法。

DESCRIPTION (provided by applicant): Children with specific language impairment (SLI), a spoken language impairment in the absence of any obvious causal factors, make up about 7% of the population and have substantially poorer long-term academic and social outcomes as compared to their peers. During the preschool years, deficits in acquisition and use of grammatical morphology and syntax are among the most commonly observed impairments. Current interventions for morphology and syntax have limited efficacy (Law et al., 2004). Enhanced language intervention techniques would improve the long-term positive outcomes for this population. This research grant draws on input-driven approaches to typical language learning, which propose that highly frequent examples help scaffold acquisition of grammatical constructions by providing a stable exemplar to facilitate linking the meaning of grammatical construction to the form. Similarly, theories of word learning and learning of artificial grammars suggest that using atypical exemplars may promote learning to a greater degree as compared to highly typical exemplars. The research program extends these theories to intervention by using computational modeling and behavioral methods for the purpose of examining the role of stimulus selection and presentation on intervention outcomes. Two early- efficacy studies will be carried out. One is focused on teaching children with SLI to use past tense using two different means of selecting verbs for use in intervention. The second is focused on teaching a novel morpheme to children with SLI and children with typical language using three different input distributions. The results of the behavioral studies will be linked to computational models enabling us to test input structures and means of selecting stimuli that go beyond what is feasible to examine behaviorally. This will make it possible to test hypotheses about what will improve interventions more efficiently than with traditional behavioral methods alone. The investigator has substantial experience working with children with SLI and assessing the use of grammatical morphology in children. However she lacks expertise in the areas of computational modeling, design of intervention studies, and communicating results to speech pathologists. The training component of this grant will ensure that the investigator is capable of using computational modeling techniques in concert with behavioral results to better understand how children learn. It will also enhance the investigator's ability to independently carry out intervention studies and communicate research results to practitioners. The University of Iowa is well known for work in the areas of specific language impairment and computational modeling and has a long history of clinically relevant research. Extant collaborations between researchers in psychology and communications disorders make this an ideal environment for this training to take place. The training opportunities will help the investigator launch an independent research career that investigates ways to improve remedi-ation of language disorders.