Neural Dynamics of Spoken Word Recognition

项目摘要/摘要 大脑和认知科学的研究提供了压倒性的证据，即使是 像理解口语这样看似简单的任务依赖于 各种不同类型的信息(例如听觉的、语音的、语音的、词汇的、 语义等)。这些互动是值得注意的，因为它们似乎允许听众， 包括许多中风患者的脑部局灶性损伤 表示这些类型的信息，以识别口语，即使在 由于清晰度差，语音合成不完善，噪音大，语音不清 环境，或由于数字减少或滤波而导致的信号质量差(例如，当 通过人工耳蜗植入、助听器或电话连接不良而产生的耳鸣)。的确有 仍在激烈辩论其中许多互动的性质或必要性，原因是 当前行为成像和单峰成像可解释性的内在局限性 范例。该方案通过将MRI、脑磁图和EEG相结合来解决这些限制 提供高时空分辨率的脑活动演变图像的数据 言语感知任务。这些数据将被提交给格兰杰因果分析， 这使得研究人员能够直接检查 与加工相关的不同脑区激活的关系 特定类型的信息。使用这些技术，拟议的研究 考察了导致以下现象的机制：(1)语音的离散分类 声音(范畴知觉)，(2)频率或音位对知觉的影响 (3)语义语境对言语感知的影响。拥有 确定了本地化功能如何相互作用以产生稳健的语音感知 未受损的监听者，然后将使用这些工具来检查如何保存 18例单侧局灶性脑损伤后的过程重组和整合 中风后几个月，允许失语症患者恢复功能。这些数据 解决以前无法解决的关于大脑功能的核心问题， 人类的言语感知和支持失语症恢复的机制 病人。作为基础研究，这项工作应该对 局灶性损伤患者的研究、评估和康复。

Project Summary/Abstract Research in the brain and cognitive sciences provides overwhelming evidence that even seemingly simple tasks like understanding spoken words rely on the interaction of a variety different types of information (e.g. auditory, phonetic, phonological, lexical, semantic etc.). These interactions are remarkable in that they appear to allow listeners, including many stroke patients with focal damage to brain regions that are responsible representing these types of information, to recognize spoken language, even when speech is unclear due to poor articulation, imperfect speech synthesis, noisy environments, or poor signal quality due to digital reduction or filtering (e.g. when experienced via cochlear implant, hearing aids, or a poor phone connection). There is still vigorous debate about the nature or necessity of many of these interactions due to intrinsic limitations in the interpretability of current behavioral and unimodal imaging paradigms. This proposal addresses these limitations by integrating MRI, MEG and EEG data to provide high spatiotemporal resolution images of evolving brain activation during speech perception tasks. These data will be submitted to Granger causality analysis, which allows researchers to directly examine patterns of cause and effect in the relationship between activation of different brain regions associated with the processing of specific types of information. Using these techniques, the proposed research examines the mechanisms that give rise to: (1) the discrete categorization of speech sounds (categorical perception), (2) frequency or phonotactic effects on the perception of speech sounds, (3) the influence of semantic context on speech perception. Having identified how localized functions interact to produce robust speech perception in unimpaired listeners, these tools will then be turned to examine how preserved processes are reorganized and integrated after unilateral focal brain damage in the 18 months following stroke, to allow the recovery of function in aphasic patients. These data address central, previously irresolvable questions about brain function, the robustness of human speech perception, and the mechanisms that support recovery in aphasic patients. As basic research, the work should have wide-reaching implications for the study, assessment, and rehabilitation of patients with focal damage.