多模态PET/MRI脑成像联合处理方法研究

重大脑疾病，如阿尔茨海默病、抑郁症等，已经成为威胁人类健康甚至生命的主要疾病。正电子发射断层成像（PET）、磁共振成像（MRI）等无创的神经影像技术为脑疾病的研究提供了有效的手段。不同成像模态能够提供不同的脑功能或结构信息。利用PET/MRI多模态成像对脑疾病进行深入全面的研究已成为目前研究的重点。目前常用的多模态影像数据处理方法是先对各个模态分别处理，然后再对结果进行汇总及分析。虽然这种方法利用了多种模态提供的不同信息，但没有考虑不同模态之间的关系及相互影响。因此，建立多模态脑影像数据联合分析方法已成为目前亟需解决的问题。本项目拟基于独立成分分析（ICA）建立适合多模态PET/MRI脑影像数据联合分析的方法，包括（1）建立适合三维脑影像的非负ICA分解算法；（2）建立多模态脑影像ICA分析流程；（3）使用模拟数据对所建立的分析方法进行评估，并将其用于阿尔茨海默病的PET/MRI数据分析。

Brain diseases have become a major killer of human life, such as Alzheimer, Depress and so on. Neuroimaging, such as positron emission tomography (PET), and magnetic resonance imaging (MRI) have been widely used in the clinical diagnosis and basic research of brain diseases. The brain functional activity or structural information could be obtained by multi-modality neuroimaging. Based on multi-modality PET/MRI neuroimaging techniques, brain researches would be in-depth and comprehensive. Data analysis is important in multi-modality neuroimaging studies. Firstly, the multi-modality neuroimaging data is analyzed separately and numerous analysis results could be obtained. For example, the FDG-PET functional images were analyzed by standardized uptake value ratio based on regions of interests, while the MRI structural images are analyzed by voxel-based morphology. Afterwards, these analysis results are combined and analyzed manually, which is superficially and subjectively. Until now, how to establish a method of combine multi-modality data analysis combin is still an unresolved problem. Therefore, this project will construct an efficient data analysis method based on independent component analysis (ICA) method for multi-modality PET/MRI images, comprising (1) to construct a developed non-negative ICA decomposition algorithm; (2) to construct a standard ICA analysis process for multi-modality PET/MRI images; (3) to evaluate the constructed method based on simulated data and applicate it in multi-modality PET/MRI data analysis of Alzheimer disease.