This study aims to advance the understanding of mental disorders and propose novel therapeutic approaches by developing a mesoscopic-level digital brain model. Based on predictive processing theory, the model integrates exteroceptive, proprioceptive, and interoceptive signals to simulate pathological states and explore mechanisms underlying symptom formation of psychiatric disorders. Additionally, intracranial EEG and autonomic nervous system data will be used to construct precision digital brain models that reflect individual-specific neural characteristics. Furthermore, the study will employ XR technology to develop "NeuroVR," a platform that replicates perceptual alterations such as hallucinations in hyper-realistic detail, using phenomenological analysis to refine the model. Finally, these advancements will inform the exploration of new therapeutic interventions leveraging neural modulation and XR technology.