Major depressive disorder (MDD) is a complex and severe mental illness, characterized by multiple symptoms, and is a leading cause of non-fatal health loss. Despite this, approximately 30% of patients do not respond to standard pharmacological or psychological treatments. Currently, we lack objective brain-based biomarkers to differentiate between natural mood fluctuations and situations requiring intervention. To address this issue, we employed a novel electrophysiology recording device and applied deep brain stimulation (DBS) to 12 MDD patients. Our study aims to use long-term invasive neural signal collection and machine learning techniques to reveal the complex relationship between these signals and depressive symptoms. By applying advanced machine learning algorithms, our goal is to establish highly accurate prediction models to identify biomarkers associated with the occurrence and progression of depression. The research will focus on the spatiotemporal features of neural signals and build personalized depression decoding models based on individual differences through the integration and analysis of large-scale data. By delving into the information contained in neural signals, we hope to contribute to the development of personalized treatment approaches for depression.