基于同步EEG-fMRI采集的情绪认知重评数据特征融合分析研究

发布时间：2018-01-14 12:45

本文关键词：基于同步EEG-fMRI采集的情绪认知重评数据特征融合分析研究　出处：《自动化学报》2016年05期 　论文类型：期刊论文

【摘要】：脑电(Electroencephalography,EEG)与功能磁共振成像(Functional magnetic resonance imaging,f MRI)为脑科学研究提供了互补的时空信息.为研究大脑在对情绪图片采取认知重评策略时的神经活动,基于同步采集的EEG-f MRI数据,应用典型相关分析、经验模态分解及k-均值聚类等算法对融合情绪数据进行交叉关联和盲源分离,得到空间上的f MRI图像和与之对应的EEG时间演变信号.结果表明:时域上,CCA分离出的脑电成分在认知重评状态下有明显的晚期正电位(Late positive potential,LPP)(潜伏期200 ms~900 ms)出现,而且认知重评策略诱发下的LPP波幅明显小于观看负性诱发的LPP波幅(F(1,224)=28.72,P0.01),而大于观看中性诱发的LPP波幅(F(1,224)=63.32,P0.01);与之对应的空域上,可以明显地看出和情绪调节相关的扣带回,额叶、颞叶等区域有明显激活区,采用情绪认知重评策略时的脑区激活强度明显小于观看负性状态,而大于观看中性,且观看中性状态下被激活的与情绪相关的区域相对较少.研究表明,这种融合数据分析技术通过计算两种模态数据之间潜在的线性相关性,可以有效地分离出大脑在时空上神经活动情况,达到了同时描绘出大脑神经活动的时间信息与空间信息的效果.
[Abstract]:Electroencephalography. EGG and functional magnetic resonance imaging. FMRI provides complementary temporal and spatial information for brain science research, and is based on synchronous EEG-f MRI data to study the neural activity of the brain when it adopts cognitive reassessment strategies for emotional images. Using canonical correlation analysis, empirical mode decomposition and K-means clustering algorithms, the fusion emotional data are cross-correlated and blind source separation. The spatial f MRI image and the corresponding EEG time evolution signal are obtained. The results show that: in time domain. The EEG components isolated by CCA had significant late positive positive potential under cognitive reassessment. LPPN (latent period 200 ms~900 Ms) appeared, and the amplitude of LPP induced by cognitive reappraisal strategy was significantly lower than that of LPP induced by negative viewing. The results showed that there was a significant difference between the two groups (P 0.01), while the amplitude of LPP was higher than that of the neutral LPP (63.32). In the corresponding airspace, the cingulate gyrus, frontal lobe, temporal lobe and other regions associated with emotional regulation can be seen to have obvious activation areas. The activation intensity of the brain area was significantly lower than that of the watching negative state, but greater than that of the viewing neutral, and there were relatively few emotion-related areas activated in the viewing neutral state. The results showed that the activation intensity of the brain area was significantly lower than that of the watching neutral state. By calculating the potential linear correlation between the two modal data, this fusion data analysis technique can effectively separate the brain neural activity in time and space. At the same time, it can describe the temporal and spatial information of brain nerve activity.
【作者单位】：常州大学信息科学与工程学院;南京医科大学附属常州市第二人民医院影像科;南京大学社会学院心理系;
【基金】：国家自然科学基金项目(61201096;51307010) 常州市科技项目(CE20145055) 江苏省青蓝工程资助~~
【分类号】：TP311.13
【正文快照】： 引用格式邹凌,严永,杨彪,李文杰,潘昌杰,周仁来.基于同步EEG-f MRI采集的情绪认知重评数据特征融合分析研究.自动化学报,2016,42(5):771-781Feature Fusion Analysis of Simultaneously Recorded EEG-f MRI inEmotion Cognitive ReappraisalZOU Ling1YAN Yong1YANG Biao1LI We

【相似文献】