A research paper by scientists at Fudan University presented BioPINN-LM, a hybrid framework that combines physics-informed neural networks and multimodal large language models to rapidly predict aortic aneurysm wall stress and generate interpretable biomechanical risk assessments.

With the advancement of long-term crewed deep-space exploration missions, the effects of the microgravity environment on astronauts' brain function have become an important topic in aerospace medicine. The disappearance of the hydrostatic pressure gradient during spaceflight induces cephalad fluid shifts, which can alter cerebral blood flow (CBF) and are closely associated with visual impairment, spatial disorientation, and cognitive decline. However, existing studies have predominantly focused on the steady-state phase during the middle and late stages of microgravity exposure, with insufficient understanding of the characteristics of CBF fluctuations during the early phase of postural transitions. Moreover, as an indicator of terminal perfusion in the cerebral microcirculation, the coupling relationship between CBF and the upstream hemodynamics of the internal carotid and vertebral arteries has not been systematically investigated. In addition, magnetic resonance imaging (MRI) cannot be applied in real time during spaceflight due to its bulky equipment, whereas ultrasound offers the advantages of portability, real-time capability, and radiation-free operation. Therefore, elucidating the association between early macrovascular hemodynamic changes under simulated microgravity and regional CBF alterations, and evaluating the feasibility of using ultrasound parameters as portable predictors of CBF, holds significant scientific importance and application value.