It has been reported that patients with a single MEG spike cluster have better postoperative outcomes compared to patients with multiple clusters. However, the problem of identifying a genuine source in the multiple MEG spike clusters remains unsolved. The aim of this study was to determine the focus of the epileptic activity in a situation with multiple MEG spike clusters. Five patients with intractable epilepsy with multiple MEG spike clusters were retrospectively investigated. They had pathologically proven type IA focal cortical dysplasia (FCD) with discordant multimodal presurgical evaluations, and showed a favorable postoperative outcome. MEG spikes were localized and clustered using a hierarchical clustering method. Then, effective connectivity (a phase-slope index method) referring to the causal interaction between distant structures in the brain was applied to the source waveforms extracted from the multiple MEG spike clusters. Finally, the information source, that is, the driver region between multiple clusters, could be identified. We found that 4 of the 5 FCD patients exhibited the driver regions were coincident with the resection area, which was also quite consistent at different epochs. Our results suggest that effective connectivity analysis has a potential value as a presurgical evaluation when multiple MEG spike clusters are found.