基于灰建模的同步EEG-fMRI脑功能动态建模与识别

项目名称： 基于灰建模的同步EEG-fMRI脑功能动态建模与识别

项目编号： No.61273250

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 谢松云

作者单位： 西北工业大学

项目金额： 80万元

中文摘要： 为克服现有技术无法同时在时间和空间上得到脑功能的精确表述这一缺陷，提出脑功能动态建模和识别新方法，其关键在于高效快速的建模方法以及快速精准的多模态信息融合技术。本项目提出将大脑视为灰色系统，研究采用灰建模技术建立脑功能动态模型，将模型参数作为特征量进行脑功能状态识别，计算量小、准确度高，可实现脑功能动态模型快速精确地建立；而建模的前提条件是获得同时在时间和空间上都具有高分辨率的脑功能融合图像，因此，提出从同步检测的EEG-fMRI多模态脑功能图像着手，研究采用NURBS曲面建模技术，在fMRI约束条件下进行EEG三维成像，实现脑功能信息在时间和空间上的融合，显著提高脑功能动态模型的空间分辨率和动态特性，为灰建模提供数据条件。新方法为研究脑功能动态特征建立完整的理论依据，为进一步理解和揭示脑功能状态的全过程及认知规律奠定重要理论基础，对心理状态监测、脑-机接口等研究有重要意义和应用前景。

中文关键词： 同步脑电-功能核磁共振信息；多模态融合；脑功能动态建模；灰理论；特征提取与识别

英文摘要： To overcome the backage of less accurate presentation of brain function on the space and time simultaneously, a new method of dynamic modeling and recognition of brain function is proposed. The key techniques for dynamic modeling of brain function are high efficiency modeling method and accurate information fusion technique which both of them should be with very high speed. We take brain as a grey system and put forward to adopt grey modeling technique for dynamic modeling of brain function, which is accurate with less computation time. The parameters of grey model are taken as the features for recognition. Before building the grey model we need to obtain the fusion image of simultaneous EEG-fMRI to provide dataset for modeling. Inorder to fuse the multi-modal signals of simultaneous EEG-fMRI we adopt the NURBS modling technique to build accurate model of the head based on structural MRI. Then we obtain the fusion image through building a kind of 3-D EEG image of head under the restrain of fMRI based on the head model. Both of the space resolusion and characterictics of dynamics of the model of brain function will be improved by the new method. Our research will set up an integrated theory for research on dynamic feature of brain function and lays a theoretical foundation for further understanding and explaining

英文关键词： Simultaneous EEG-fMRI Information；Multimodal Fusion；Brain Function Dynamic Modeling；Grey Theory；Feature Extraction and Recognition