Approved Research

Heterogeneous Graph Neural Networks Research For Brain Network Analysis and Disorder Diagnosis

Lay summary

Our research project aims to develop a powerful tool called Heterogeneous Graph Neural Networks (HGNNs) to better understand the human brain and improve the diagnosis of brain disorders. The rationale behind this project is that the human brain is incredibly complex, and understanding its intricate networks is crucial for diagnosing and treating neurological disorders. HGNNs allow us to analyze different types of brain data, such as structural, functional, and genetic information, all at once. This holistic approach gives us a comprehensive understanding of how the brain works and how it is affected by disorders. The project duration is expected to be approximately three years. In the first stage, we will process brain images and extract important information from different brain regions. Then, we will build heterogeneous graphs that capture the connections between these brain regions. Next, we will develop methods to extract high-order features from the brain data, allowing us to identify patterns and biomarkers associated with specific disorders. Finally, we will analyze the brain networks using the HGNNs and develop new diagnostic methods. The impact of this research on public health is significant. By using HGNNs, we can potentially identify brain abnormalities and diagnose disorders at an earlier stage, leading to timely treatments and improved outcomes for patients. This research may also uncover new insights into the mechanisms of brain disorders, leading to the development of more effective therapies. Ultimately, our goal is to revolutionize brain disorder diagnosis and improve the lives of individuals affected by these conditions.