Denoising-Assisted Rapid Multiphoton Imaging for Analysing Traumatic Penumbra Microenvironment in Rat Brain

Accurate identification of the traumatic penumbra (TP) is essential for understanding secondary injury after traumatic brain injury. However, conventional imaging cannot capture microenvironmental alterations at subcellular resolution. In this study, denoising-assisted multiphoton microscopy (MPM) was used to investigate the microstructural features of the TP in rat brain tissue. Image processing and quantitative analysis were performed to characterize normal tissue, penumbra, and core regions. To accelerate high-quality imaging, we proposed a Dual-Branch Collaborative Network (DBCNet) to restore low-noise images from single-scan data using multi-scan references. The results show that MPM reveals key pathological features such as intracellular edema, vasogenic edema, and cytoplasmic matrix disruption. DBCNet effectively suppresses noise, enhances fine structures, and reduces imaging time to one-fourth without hardware modification. These findings demonstrate that denoising-assisted MPM provides a rapid and efficient strategy for high-resolution imaging and quantitative assessment of the TP microenvironment.