OBJECTIVE : Depression after stroke (DAS) is a serious complication of stroke that significantly restricts rehabilitation. Brain imaging technology is an important method for studying the emotional network of DAS. However, few studies have focused on dynamic interactions within the network. The aim of this study was to investigate the emotional network of frontal lobe DAS using the multivariate Granger causality analysis (GCA) method, a technique that can estimate the association among the brain areas to analyze functional magnetic resonance imaging (fMRI) data collected from DAS and no depression after stroke (NDAS). METHODS : Thirty-six first-time ischemic right frontal lobe stroke patients underwent resting-state fMRI (rs-fMRI) scans. The clinical assessment scale used for screening subjects was as follows: the 24-item Hamilton Rating Scale for Depression (HAMD-24), the National Institutes of Health Stroke Scale (NIHSS), the Mini-Mental State Examination (MMSE), and the Barthel Index (BI). The multivariate GCA method was used to analyze fMRI data collected from DAS and NDAS. RESULTS : The results showed positive regulations in the order from the ventromedial prefrontal cortex (VMPFC), the anterior cingulate cortex (ACC), and the amygdala (AMYG) to the thalamus, and when the interaction order is opposite, the moderating effect is negative. The thalamus could predict the negative activity of the insular (IC) via the ACC. The dorsolateral prefrontal cortex (DLPFC) could predict the activity of the ACC via the temporal pole (TP). CONCLUSIONS : This study found a VMPFC-ACC-AMYG-thalamus emotional circuit to explain the network between different brain regions associated with DAS. The DLPFC and TP play an important role in the emotional regulation of DAS, and the function of the IC is regulated negatively by the thalamus. These findings advance the neural theory of DAS, which is based on the functional relationship between different brain areas.