多模态言语运动评估用于帕金森症构音障碍机制研究

Although the primary manifestations of Parkinson’s disease (PD) such as tremor, rigidity, and akinesia have been found to respond well to dopamine therapy, some of the most troublesome movement disorders associated with speech and postural control, or the so-called “axial signs”, remain largely unresponsive. This may suggest that a distinct basal ganglia (BG) pathways/neurotransmitter systems subserve mechanisms of speech production. A vast body of literature exists on different analytic approaches of speech production. However, speech deficits in PD tend to vary according to the specific speaking tasks and resulted in problematic articulatory movement. Results from previous studies may not be directly comparable due to the use of different measuring yardsticks for assessing motor speech functions of PD patients. The lack of universal, integral measurements for motor speech functions does not allow a comprehensive, quantitative, and qualitative description of speech production of individuals with PD..The present project will attempt to answer this question by examining motor speech subsystems using a range of behaviors from vowel prolongation to reading. The measurement approach will utilize an elegant, innovative assessment protocol, which include a series of noninvasive, simultaneous transduction of kinematic and acoustic measurements, to span the observation of multiple levels of speech production (e.g., respiration, phonation, articulation, and prosody) as well as insuring relevance to motor speech function. These measurements will be obtained from a relatively larger group of Chinese PD patients (n=20) from Shenzhen, which allows the cross-cultural and cross-linguistic comparisons. Results will help us develop and verify the new model in the understanding of neurological mechanisms of dysarthria in PD patients. As such, a novel multimodal assess system will be developed. The system will integrate the acoustic measurements and kinematic measurement by using audio recorder and facial capture system. The features will be selected with maximal joint relevance to UPDRS score according to fusion of hypothesis-driven model and data-driven model. Kinematic and acoustic dataset to be built in present project will be used to test the reliablity and robust of proposed system. Findings of the proposed study will also help design new rehabilitation management model and provide rehabilitation feedbacks for populations with movement disorders, such as patients with PD, and as a result to improve their quality of life..

帕金森症构音障碍是帕金森症的核心症状之一，严重影响病人的情感、交流和日常生活。药物、手术治疗对帕金森症患者的言语困难改善的效果不佳。无法改善干预治疗方案的原因是当前对基底神经节功能以及其对言语运动系统的病理影响机制尚不能完全解释。本研究将致力于设计一系列创新性、多模态的言语运动功能评估体系，运用三维电磁发音运动跟踪、面部运动捕捉以及呼吸\声学参数分析技术追踪发音器官运动特征，全面、客观地评估帕金森症对言语运动功能的影响以及多巴胺替代疗法对其改善作用。以此为基础，重构基底神经节功能模型及帕金森症构音障碍损伤模型，并明确非多巴胺能运动环路对言语运动功能的影响。最后，根据帕金森症构音障碍损伤模型及其临床表现，设计融合口面部运动和声学参数的新型帕金森症构音障碍评估系统。本研究的成果可用来指导未来的包括使用左旋多巴及言语康复治疗的干预性治疗手段的改进。

帕金森症是仅次于阿尔茨海默症的老年神经退行性疾病。据统计，我国有超过170万的帕金森病患者，其中接近90%的帕金森症患者都伴有语言和发音障碍，是帕金森症的核心症状之一，严重影响病人的情感、交流和日常生活。然而，药物 、手术治疗对帕金森症患者的言语困难改善的效果不佳。无法改善干预治疗方案的原因是当前对帕金森症以及其对言语运动系统的病理影响机制尚不能完全解释。因此，本项目在国家自然基金青年项目的支持下，围绕这一核心关键问题展开研究：首先，构建了多模态言语运动功能评估体系，运用三维电磁发音运动跟踪以及声学参数分析技术追踪发音器官运动特征，全面、客观地评估帕金森症对言语运动功能的影响以及多巴胺替代疗法对其改善作用。其次，我们通过脑电、事件相关电位以及源定位技术，对帕金森症构音障碍中听觉-言语运动反馈损伤的神经机制进行深入分析，进一步分析通过EMA数据探讨了多巴胺能对帕金森症患者的构音障碍的影响，重构帕金森症构音障碍损伤模型；最后，我们对声学和言语运动学相关特征的优化选取展开研究，包括探讨基于能量特征分解以及非线性参数对帕金森症患者构音障碍的有效性，以及敏感反映舌运动特征的相关特征参数选取，构建了嗓音质量客观评估模型，可以有效的、自动对病理嗓音质量进行客观分级，从而为开发可用于汉语病理嗓音障碍程度自动分级评估系统，使之能为嗓音病理提供快速、非侵入性的辅助诊断依据打下良好的研究基础。本研究的成果可用来指导未来的包括使用左旋多巴及言语康复治疗的干预性治疗手段的改进。