EEG源成像技术难以准确估计脑活动位置和尺寸，而fMRI技术时间分辨率过低。结合EEG、fMRI的多模态脑功能成像方法，利用二者的互补特性，能更准确地重构脑激活区的位置、尺寸和时间过程，为脑疾病诊断和认知科学研究提供技术保障。本项目拟在前期EEG弥散源成像工作基础上，利用贝叶斯概率框架和凸分析技术，充分挖掘EEG-fMRI包含的时空信息，建立基于时空约束的EEG-fMRI脑功能成像算法框架，开发脑源信号的高效重构算法，实现精确估计脑活动时空特性，并应用于脑功能网络和临床数据分析。拟主要研究以下问题：1）进一步拓展基于时空约束的EEG源成像算法；2）融合EEG-fMRI信息的多模态脑功能成像算法；3）基于所开发脑功能成像算法的脑网络分析。以上算法将在实验和临床EEG和fMRI数据上进行验证改进，为提高脑功能成像性能提供新方法和技术。

It is difficult for EEG source imaging methods in estimating the locations and extents of brain active areas accurately, while fMRI suffers from its low temporal resolution. The multimodal brain functional imaging methods combining EEG and fMRI, using the complementary features between EEG-fMRI, are able to estimate the locations, extents and time courses of brain activities more accurately, which provides technological support for brain diseases diagnoses and cognitive science studies. Based on our prior work of EEG extended source imaging, the objective of this project is to establish a framework for brain functional imaging methods using EEG-fMRI based on spatio-temporal constraints, and develop efficient algorithms to reconstruct brain source signals, using Bayesian probabilistic framework and convex analysis, which fully exploit the spatio-temporal information of EEG-fMRI. Thus we can estimate the spatio-temporal properties of brain activities accurately. These imaging results could be used to analyze brain functional network and clinical data sets. The following important topics will be studied in this project: 1) EEG source imaging methods based on spatio-temporal constraints; 2) multimodal brain functional imaging algorithms based on EEG-fMRI fusion; 3) brain functional network analysis based on the proposed brain functional imaging algorithms. All of the proposed algorithms will be verified on experimental and clinical data sets, providing new algorithms and technologies to improve the performance of brain functional imaging.