Academic Report of School of Mathematical Sciences [2026] No. 093

(Series Report for High-Level University Construction No. 1352)

Title：From Physical Inversion to Neural Inversion: Magnetic Resonance Imaging, Functional Decoding, and Alignment of Large Model Representations

Speaker：Professor Cui Zhuoxu(Shenzhen Institutes of Advanced Technology，Chinese Academy of Sciences)

Time：9:00-11:00, Sept. 24, 2026

Location：Room 633, Huiyan Building, Yuehai Campus, Shenzhen University

Abstract: Inverse problems are a common core element in both magnetic resonance imaging (MRI) and neural decoding. Traditional MRI reconstructs images from k-space signals, and its forward model is primarily determined by physical encoding processes, which are typically relatively well-defined; in contrast, fMRI visual decoding attempts to infer external stimuli from brain activity signals. While this also falls under the category of inverse problems, its key forward process—the mapping from visual stimuli to neural activity—is not fully analyzable. Consequently, the core challenge of neural decoding has shifted from “image inversion under a known physical model” to “stimulus inversion under an unknown neural model.” This presentation will take inverse problems as its central theme, beginning with MRI reconstruction and transitioning to fMRI visual decoding. It will further discuss whether large language models can serve as computable approximations of visual representations in the human brain and introduce relevant research progress centered on the question: “Can representations from large language models be aligned with human neural activity?”

Speaker Profile：Cui Zhuoxu, Research Fellow at the Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences. His primary research areas include magnetic resonance imaging and medical artificial intelligence. Over the past five years, he has published more than 50 papers in academic journals and conference proceedings. He has served as principal investigator on grants from the National Natural Science Foundation of China, the Shenzhen Natural Science Foundation, and the Chinese Academy of Sciences Cluster Program, among others; participated in major projects such as the Chinese Academy of Sciences Pioneer Program B and the National Key Research and Development Program; and was selected for the Chinese Academy of Sciences Stable Support Program for Young Research Teams.

Faculty and students are welcome to attend!

Invited by: Sun Xiaoli

School of Mathematical Sciences

September 21, 2026