Professor Du Yuhui's team has long focused on interdisciplinary research at the intersection of brain science and artificial intelligence (AI). Integrating knowledge from multiple disciplines such as computer science, medicine, and neuroscience, the team has originally developed cutting-edge international image processing and AI technologies for mental illness diagnosis. These technologies not only assist in the accurate diagnosis of mental illnesses but also deeply reveal their neural mechanisms. Meanwhile, the team is committed to using brain imaging to explore the laws of brain operation, development, and aging, promoting the development of brain-inspired intelligence, facilitating the in-depth integration of AI and brain science, and realizing the efficient transformation of basic research achievements into clinical diagnosis and treatment.

The research focuses on the core scientific issues of multimodal brain imaging in mental illness diagnosis. Aiming at technical challenges such as high dimensionality, strong noise, small sample size, and modal differences of brain imaging data, as well as difficulties in clinical diagnosis of mental illnesses (including subjectivity, individual differences, inaccurate labels, and easy confusion between diseases), the team has proposed a series of methodological innovations. These mainly include: accurate extraction of static and dynamic brain functional networks; efficient feature selection and extraction to support the discovery of key sparse biomarkers; multimodal fusion-based brain disease classification methods; semi-supervised and noisy label learning strategies suitable for mental illnesses; identification of disease biological subtypes based on brain imaging; and construction of foundation models for brain functional networks.