基于elastic net方法的静息态脑功能超网络构建优化

脑网络分析已广泛应用于神经影像领域的研究。超网络构建方法被提出用于描述多个脑区之间的高阶关系。超网络是根据静息态功能磁共振成像时间序列通过稀疏线性回归方法构建。在已有文献中，用于构建超网络的稀疏线性回归模型是采用lasso方法解决。然而这种方法存在局限，在超边构建时不能够有效的解决脑区之间的组效应。针对这一问题，本文提出将elastic net方法引入到超网络构建中,并且应用于抑郁症患者与正常被试的分类。实验结果显示基于lasso与基于elastic net的方法分别可以达到83.33%与86.36%的分类准确率。分类结果表明与原有方法相比,基于elastic net的方法可以得到更为有效的特征以及更好的分类效果。

The optimization of resting-state brain functional hyper-network construction based on elastic net

Brain network analysis has been widely applied to the studies of neuroimaging field. Hypernetwork construction method was proposed to characterize the high-order relationships among multiple brain regions. Brain functional hyper-network is constructed by the sparse linear regression method according to the resting state functional magnetic resonance imaging time series. In the literature, the sparse linear regression model for hyper-network is solved by lasso method. However, this method has a limitation which cannot solve the grouping effect effectively among brain regions when constructing hyperedges. To solve this problem, the article proposed the introduction of the elastic net method into the hyper-network construction, and the method was applied to the classification of depression patients and normal controls. Experimental results showed that the lasso-based method and elastic net-based method can achieve 83.33% and 86.36% of classification accuracy, respectively. The classification results indicate that compared with the original method, the elastic net-based method could obtain more effective features and better classification results.