Visualizing microvessels in the brain with ultrasound…DGIST develops ‘ULM-Lite’ technology

□ The DGIST (President Kunwoo Lee) research team, led by professors Jaesok Yu and Jungho Hyun of the Department of Robotics and Mechatronics Engineering (with doctoral students Hyojin Seong and Jinhwan Jeong as lead authors), improved blood vessel observation technology using ultrasound. The team developed ‘ULM-Lite,’ which significantly boosted the efficiency of ultrasound localization microscopy (ULM), an ultrasound-based super-resolution imaging technology, paving the way for viewing microvessels with small data.

□ The ultrasound commonly used in hospitals is useful for observing the shape and movement of organs inside the body. However, it has limitations in distinguishing microvessels, which are thinner than hair. ULM is a technology that was developed to overcome these limitations. It tracks the ultrasound contrast agent (microbubbles) in the blood as it moves, tracing each one individually to reconstruct the microvessel structure in super-resolution. By analyzing the trajectories of these movement points, a map of blood vessels at a scale of a few tenths the thickness of a hair—completely invisible with conventional ultrasound—can be created.

□ The issue is that this technology is too heavy. ULM captures thousands of ultrasound images at very high speeds and tracks the signals of countless microbubbles moving within them. This process produces several gigabytes (GB) of data every second. It is similar to creating a movie by drawing each frame by hand. While the results can be very detailed, it is also a highly time-consuming and resource-intensive process. This has made it difficult to use conventional ULM effectively in real-world medical settings, where long-term studies or immediate results are often needed.

□ To overcome the limitations, the research team of Professors Jaesok Yu and Jungho Hyun developed a new analysis method that retains only essential information and significantly reduces unnecessary data from ultrasound signals. The team created the ‘ULM-Lite,’ which lowers the ‘effective bandwidth’ of signals to about 67% and efficiently extracts only the information necessary for mapping blood vessel structures. As a result, it maintains image clarity while reducing data size and greatly increasing processing speeds.

□ The core of this research is that it significantly reduces data usage while utilizing existing ultrasound equipment. 'ULM Lite' cuts data volume by about one-third by preserving only the essential information from the ultrasound signal, while nearly maintaining the same image quality. It can be implemented without replacing equipment and boosts image processing speeds by roughly 30%. Additionally, it clearly visualizes the entire brain non-invasively without requiring surgery or fluorescent materials, thereby, greatly improving the efficiency of brain research and disease diagnosis. It shows strong potential as a tool for brain stimulation therapy and monitoring behavioral changes.

□ Professor Jaesok Yu stated, "We expect that this technology will be utilized to diagnose and treat various brain diseases, in combination with the non-invasive ultrasound brain stimulation technology that is currently under development."

□ This research was funded by the Excellent New Researcher Program and the Global Lab, supported by the National Research Foundation of Korea (NRF), the Ministry of Science and ICT, and the Ministry of Education, and it was published in the world-renowned academic journal in the field of medical ultrasound, 『Ultrasonics』.