基于功能磁共振的语言区术前定位的健康人研究

发布时间：2018-03-04 06:04

本文选题：语言功能定位　切入点：功能连接　出处：《杭州师范大学》2015年硕士论文　论文类型：学位论文

【摘要】：肿瘤病人运动、语言等重要功能区的术前定位可降低手术风险,提高病人生活质量。功能磁共振(fMRI)作为一种非侵入型的成像手段,正逐渐应用于肿瘤病人术前功能定位。基于组块设计的任务态fMRI已经作为常规技术,静息态fMRI是新兴起的技术。本文利用一种新的fMRI实验范式--基于"状态"的fMRI(即慢速事件相关设计,以下称“新范式”)定位语言功能区,试图比较新范式及其相应的数据分析方法和传统技术的差别,为更加准确的术前定位提供参考。为此,我们收集了51名健康被试磁共振扫描的数据。被试分别进行了组块设计的任务态语言fMRI,新范式下的语言fMRI和静息态fMRI扫描。对比新范式和静息态fMRI的局部功能连接(采用局部一致性指标刻画,ReHo)分析结果,并对比两种范式下种子点功能连接(FC)和独立成分分析(ICA)得到的语言功能网络的结果,发现新范式下的语言区的局部功能连接相比静息态结果有明显上升,证明了新范式下的语言任务参与度更好;我们还发现新范式fMRI数据的FC语言定位的敏感性和特异性比静息态更高,而对于数据驱动的ICA方法来说二者的定位结果无显著差异。针对ICA方法,本研究对比了每个被试群组ICA的个体水平重建结果,初步证明了语言区可以通过静息态fMRI结合ICA得以定位,且定位结果和新范式下ICA定位结果,以及语言区解剖位置高度一致。本研究还对比了基于"状态"的fMRI的语言定位结果和传统任务态fMRI的语言任务激活结果,发现基于"状态"的范式在语言定位中的总体表现优于传统任务范式。具体来说,静息态和新范式下,语言区的定位结果的特异性高,定位结果基本一致,可以相互验证;而任务激活结果特异性低,很多非语言特异的脑区一并激活,为结果的理解带来了难度。本研究的结果说明在语言功能区的定位上,两种基于“状态”的研究范式的定位结果是可比的,而且慢速事件相关设计结合FC或ICA分析可以得到较好的定位结果,这可以作为任务态fMRI语言功能定位的有效补充。本研究的创新性在于,在基于fMRI的术前定位研究中,首次使用了慢速ER设计这种基于状态的实验范式,该范式可控性高,对病人要求不高,FC定位结果优于传统组块设计fMRI定位结果;新范式下ICA定位结果和静息态范式下ICA定位结果可比,有望在将来应用于肿瘤病人的术前功能区定位。本研究还首次证明了静息态范式下利用ICA得到的语言功能网络在执行语言任务时和语言功能相关的脑区在空间模式上高度一致,这有力的证明了利用静息态fMRI和ICA对无法完成语言任务的被试进行语言功能定位的可能性。
[Abstract]:Preoperative localization of major functional areas such as motion and language in cancer patients can reduce the risk of surgery and improve the quality of life of the patients. Functional magnetic resonance imaging (fMRI) is a non-invasive imaging method. Task-based fMRI, based on block design, has been used as a routine technique for preoperative functional localization in cancer patients. Resting fMRI is a new technique. In this paper, we use a new experimental paradigm of fMRI, which is based on "state" fMRI (slow event related design), to locate the language function area. This paper attempts to compare the differences between the new paradigm and its corresponding data analysis methods and traditional techniques in order to provide a reference for more accurate preoperative positioning. We collected data from 51 healthy subjects. The task language fMRI, the language fMRI under the new paradigm, and the resting fMRI scan were performed respectively. The local functional connections between the new paradigm and resting fMRI were compared. Then (using the local consistency index to describe the results of ReHo), By comparing the results of the two kinds of paradigms, the results of the language functional networks obtained by the seed point functional connectivity (FCC) and the independent component analysis (ICA) show that the local functional connections of the language regions under the new paradigm are significantly higher than the rest ones. It is proved that the language task participation is better under the new paradigm, and the sensitivity and specificity of FC language localization of the new paradigm fMRI data is higher than that of resting state. For the data-driven ICA method, there is no significant difference between the two methods. For the ICA method, the results of individual level reconstruction of ICA in each group were compared. It is preliminarily proved that the language region can be located by resting fMRI combined with ICA, and the localization result and the result of ICA localization under the new paradigm are obtained. The results of language localization based on "state" fMRI and the results of language task activation of traditional task-based fMRI were also compared. It is found that the overall performance of the "state-based" paradigm in language localization is superior to that of the traditional task paradigm. Specifically, under the resting state and the new paradigm, the localization results of the language area are highly specific and consistent, which can be verified by each other. However, the results of task activation are not specific, and many non-verbal specific brain regions are activated together, which makes it difficult to understand the results. The localization results of the two research paradigms based on "state" are comparable, and the slow event correlation design can be combined with FC or ICA analysis to obtain better localization results. This can be an effective supplement to the functional localization of task-based fMRI. The innovation of this study is that slow ER is used to design the state-based experimental paradigm for the first time in the pre-operation localization study based on fMRI, which is highly controllable. The result of FC localization is superior to that of traditional block design fMRI localization, and the result of ICA localization under the new paradigm is comparable to that of ICA localization under resting paradigm. It is expected to be applied to the localization of preoperative functional areas in cancer patients in the future. This study has also demonstrated for the first time that the language functional network obtained by using ICA in the resting paradigm is highly consistent with the spatial pattern of the brain regions related to language function when performing language tasks. This proves the possibility of using resting fMRI and ICA to locate the language function of the subjects who can not complete the language tasks.
【学位授予单位】：杭州师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R445.2

【相似文献】