功能MRI时间分辨率对基于格兰杰因果模型的脑效应连接的影响

目的 观察功能MRI（fMRI）时间分辨率对基于格兰杰因果关系分析（GCA）模型的脑效应连接（EC）的影响。方法 基于人类大脑连接组计划静息态fMRI（rs-fMRI）数据进行时间分辨率下采样，TR分别为0.72、1.44及2.16 s，构建3组数据。利用脑图谱将大脑划分为246个脑区，以不同阶数GCA模型分析及比较每组数据脑区间EC。结果 模型阶数1、2、3所对应的全脑显著EC在TR为2.16 s时分别为792、128及1 504条，TR为1.44 s时分别为2 932、1 920及12 500条，TR为0.72 s时分别为16 320、43 177和60 204条。结论 时间分辨率是基于GCA模型分析EC的敏感因素；fMRI数据时间分辨率越高，越易检出脑区间EC。

Impact of temporal resolution of functional MRI on brain effective connectivity based on Granger causality

Objective To observe the impact of the temporal resolution of functional MRI (fMRI) on brain effective connectivity (EC) based on Granger causality analysis (GCA). Methods Based on resting-state fMRI (rs-fMRI) in Human Connectome Project, down-sampling was performed to construct 3 sets of data with TR of 0.72, 1.44 and 2.16 s, respectively. The whole brain was divided into 246 regions according to brain atlas, and the EC of brain regions were analyzed and compared among different order GCA models. Results The amount of significant EC of the whole brain was 792, 128 and 1 504 for the model order of 1, 2 and 3 when TR was 2.16 s, respectively, which was 2 932, 1 920 and 12 500 when TR was 1.44 s, while 16 320, 43 177 and 60 204 when TR was 0.72 s, respectively. Conclusion The amount of EC based on GCA was sensitive to the temporal resolution of fMRI data. The higher the temporal resolution of fMRI data, the easier it was to detect EC of brain regions.