EEG-fMRI融合技术在静息态中的应用

Application of Simultaneous EEG-fM RI in Resting State

静息态下大脑的自发脑活动规律,对理解其对认知活动的影响以及潜在临床应用具有重要意义。同时,由于EEG与f MRI各有其缺点,对探索脑活动有各自的限制。因此,EEG-f MRI融合技术通过结合两者的优势,具有高时间和空间分辨率的特点,能更全面地研究静息态下的脑活动机制。国内外的EEG-f MRI同时技术仍处于发展阶段,加上多模态分析有其难点,例如信号噪音因人而异难以去除、融合分析方法繁多复杂、实验限制太多等,使大多研究者对其望而却步。EEGf MRI同时技术在静息态下有不同的数据处理方法,包括特殊的伪迹去除以及数据的三种融合方法。在伪迹去除方面,梯度伪迹可以通过平均伪迹模板减法(AAS)和f MRI伪迹切片模板去除(FASTR)去除,而心电伪迹则可以通过平均伪迹模板减法(AAS)、最优基础选择(OBS)和独立成分分析(ICA)去除。在数据融合方面,总结出EEG与f MRI独立对比分析、EEG-informed f MRI和f MRI-infomed EEG三种方法;其中EEG-informed f MRI是较为常用的处理方法,一般利用EEG的三种信号指标与BOLD信号做相关分析,包括功率谱(Power Spectrum)、全脑域同步(Global Field Synchronization)和微状态(Microstate)。

The spontaneous brain activity of the brain in resting state is of great significance to understand its influence on cognitive activity and potential clinical application.Also,because EEG and fM RI have respective shortcomings,they respectively have limitations on the exploration of brain activity.Therefore,simultaneous EEG-fM RI can combine the advantages of both,with high temporal and spatial resolution to more fully study the mechanism of brain activity in resting state.However,Simultaneous EEG-fM RI is still in the development stage at home and abroad and multi-modal analysis has its difficulties,such as,the signal noise varies from person to person.The integration analysis method is complex and the experiment limits too much,so that most researchers are discouraged.This article summarizes the analysis for simultaneous EEG-fM RI in resting state,including artifact removal method and three integration analysis method.Gradient artifact can be removed by means of Average Artifact subtraction(AAS)and fM RI Artifact Slice Template Removal(FASTR).And BCG artifact can be removed by means of Average Artifact subtraction(AAS),Optimal Basis Selection(OBS)and independent component analysis(ICA).In the aspect of data integration,we summarize EEG and fM RI independent analysis,EEG-informed fM RI and fM RI-infomed EEG,in which EEG-informed fM RI is a more commonly used method,and it conducts correlation analysis to EEG's three signal indexes(Power Spectrum,Global Field Synchronization,Microstate)and BOLD signal.