磁共振成像2016，Vol.7Issue(10)：769-774,6.DOI:10.12015/issn.1674-8034.2016.10.010
基于静息态功能连接的杏仁核谱聚类研究
Study on spectral clustering of amygdala using resting state functional connectivity
摘要
Abstract
Objective:According to the similarity of the functional connectivity pattern of the whole brain, the feasibility of using spectral clustering to segment the brain regions of the amygdala was studied.Materials and Methods:By resting state functional imaging according to the functional connectivity pattern of each voxel in a speciifc brain region, spectral clustering was performed to segment the amygdala region.Results:The segmentation results of subregion on the right side of the central medial nucleus (CM) were similar to that of the map, and the results of subregion on the right superficial nuclei (SF) were relatively less similar to the spectra. Cluster analysis results of the stability analysis results showed that the signal-to-noise ratio was more than 70 db, each subregion maintained 100%similarity to the results without adding noise, the pearson correlation coefifcient could be kept about 0.7 in the low signal-to-noise ratio of 10 db.Conclusion:The clustering results and JULICH maps were highly consistent, which indicated the feasibility of this method. In addition, satisfactory results were also obtained in terms of stability and anti-interference. Compared with the traditional manual segmentation method, the clustering segmentation method proposed in this paper had the advantages of simplicity, high efifciency, and high repeatability.关键词
杏仁核/磁共振成像/功能连接/谱聚类Key words
Amygdala/Magnetic resonance imaging/Functional connectivity/Spectral clustering分类
医药卫生引用本文复制引用
田洪君,高全胜,高军峰,窦顺阳..基于静息态功能连接的杏仁核谱聚类研究[J].磁共振成像,2016,7(10):769-774,6.基金项目
国家自然科学基金(编号81271659，81171385)；中国博士后科学基金资助(编号2014M552346)ACKNOWLEDGMENTS The work was supported by the National Nature Science Foundation of China (No.81271659,81171385) and the China Postdoctoral Science Foundation (No.2014M552346) （编号81271659，81171385）
