This study aims to integrate intracranial EEG and MRI data with high precision in patients with drug-resistant focal epilepsy to delineate brain networks underlying both normal functions and epilepsy. Among various brain functions, we will place particular emphasis on language. By incorporating recordings from common marmosets, we will conduct cross-species comparisons to clarify the network structures related to language. On the basis of the findings obtained from the multidimensional connectomics, we will apply mathematical modeling and machine learning to delineate the core epileptic network and to determine optimal parameters for both invasive and non-invasive stimulation. Through this approach, we seek to advance the understanding of pathophysiology of epilepsy from a network perspective and to develop novel strategies for network-based neuromodulation.