同步脑电与功能磁共振成像的单次分析方法及应用研究

The integration of signals with high spatial and temporal resolution,acquired from simultaneous electroencephalography (EEG) and functional magnetic resonance imaging (fMRI),holds a great potential for cognitive science.In this project,a novel single-trial analytic framework is proposed to find the features of and the relationship between the two modalities at the same time, which is also applied to study the neural basis for emotion regulation of the students who suffer from test anxiety.Firstly,the preprocessing robust algorithms based on single subject level are presented, which can remove the gradient artifacts, pulse related artifacts and some unknown physiological noises;Secondly,the single-trial late positive potentials, which is an effective component of cognitive reappraisal, are extracted and brought into the same range with fMRI data before entering into the fusion space; At last,the spatial maps for fMRI data and the corresponding temporal evolution for EEG data are decomposed by applying multi-set canonical correlation analysis on group level,and the co-variations across modalities from a group of subjects are obtained.Besides,the physiological processing,the composition and brain source location of test anxiety during cognitive reappraising are also discussed. This research can lead to a better understanding of human brain cognitive neural substrates of mental disorder and an alleviation of student anxiety in test as well as an improvement of human mental health level.In addition,the fusion information processing performance could be improved by understanding human brain cognitive mechanism. Therefore,this research has broad application prospects and great influences in other disciplines.

脑电与功能磁共振成像同时记录的融合信号具备较高的时间分辨率和空间分辨率，对认知科学研究有重要意义。本项目提出一套能同时获得两种模态数据特征及相互关系的单次实验分析算法，并用于研究考试焦虑者情绪调节的神经机制问题。首先，基于单个被试水平，提出能去除脑电信号中梯度噪声、脉冲相关噪声及未知生理源噪声的预处理鲁棒算法；其次，单次提取作为认知重评有效成分的晚正成分，并使其与功能磁共振成像数据同范围后进入融合数据空间；最后基于成组被试水平，采用多重集典型相关分析方法将两种模态数据分解为空域映射图和时域卷积脑电数据,以成组方式获得两者之间的共变特性，探讨考试焦虑认知重评的生理本质、构成及其神经中枢定位。本项目的研究对加深人类心理障碍的认知神经机制理解、改善考试焦虑个体的焦虑状态及提升心理健康水平有促进作用，并且借鉴人类认知机制来提高信息处理的性能，研究成果具有广阔的应用前景和辐射面。

脑电与功能磁共振成像同时记录的融合信号具备较高的时间分辨率和空间分辨率，对认知科学研究有重要意义。本项目提出对称融合和基于脑电的功能磁共振成像融合两种分析算法获得两种模态数据特征及相互关系，并用于研究认知重评的神经机制问题。主要包括以下研究内容：（1）基于单个被试水平，提出基于独立成分分析的脑电信号自动去噪算法，不仅能提高噪声检测质量，还能恢复隐藏在噪声之中的大量神经信号；（2）单次提取晚正成分波幅特征，考察认知重评对社交焦虑者/成人健康被试/学龄前儿童等不同人群情绪反应的调节，研究认知重评情绪调节的神经生理学依据；(3) 探讨功能磁共振成像数据时间序列的因果关系及有向连接的分析方法，应用密度K-means算法分析认知重评功能磁共振成像数据的脑功能连接；(4)采集同步EEG-fMRI认知重评数据，应用典型相关分析算法实现时空对称融合，计算两种模态数据之间潜在的线性相关性，同时描绘大脑神经活动的时间信息与空间信息；（5）基于单个被试水平和成组被试水平的认知重评融合数据，研究基于脑电信息的功能磁共振成像数据单次分析方法。本课题的研究成果从时间和空间两方面同时对认知重评的神经机制进行深入刻画，对理解大脑的认知功能具有重要意义。