摘要
动态因果模型(Dynamic causal modeling,DCM)是一种时空上可再生的网络模型,用来研究功能核磁共振中功能整合的因果关系,是效应连通性的分析方法,该方法是将实验设计中得到的激活区域时间序列加入到DCM模型中,实验任务的刺激响应作为对模型的扰动,利用DCM和贝叶斯估计计算出各神经元或者神经系统之间前后影响的因果关系,以及大范围的内在连接,然后利用贝叶斯因子对所设计各模型的参数做最优化选择,从中选择出符合生理的最佳模型。本文主要研究心算借位减法任务激活的左侧大脑区域,左侧顶上小叶、左侧顶下小叶和左侧额中回之间的效应连接,并得到符合生理意义的连接网络。
Dynamic causal modeling(DCM) is a spatio-temporal renewable network model. As an analytical method of causality of functional integration in fMRI, DCM is applied to study the effective connectivity. The neuroimaging time series of activated regions were put into DCM, and the trial-bound inputs were used as perturbations to the designed model. DCM was used in combination with Bayesian estimation to evaluate the intrinsic connectivity among selected neurons. Bayes factors were used to compute different neuro-physiological models with intrinsic connectivity structures, and then were used to select the optimal model. The selected regions in this mental calculation task are the left superior parietal lobule(SPL), the left inferior parietal lobule(IPL) and the left middle frontal gyrus (MFG). Finally, the connected network in conformity to physiological significance was obtained.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2009年第5期931-935,940,共6页
Journal of Biomedical Engineering
基金
国家自然科学基金资助项目(30670600)
关键词
动态因果模型
功能核磁共振
功能整合
效应连通性
贝叶斯估计
Dynamic causal modeling(DCM) Functional MRI(fMRI) Functional integration Effective connectivity Bayesian estimation