This research aims to construct an individual "digital twin brain" to elucidate mental illnesses and develop new treatments. This is a novel approach that integrates a top-down method based on brain computation theory with a bottom-up method based on biological data. We will create a personalized simulator model by incorporating individual patient's biological data, such as PET images, into an artificial neural network. Using this digital twin brain, we will conduct simulations to investigate the causes of symptoms and predict the effects of novel treatments, with the goal of developing innovative therapies.