SEOUL, September 22 (AJP) - The Korea Advanced Institute of Science & Technology (KAIST) has announced a new way to detect and treat Parkinson's disease earlier, using artificial intelligence and light to study and control brain signals.
The work was led by Professor Heo Won-do of KAIST's Department of Biological Sciences. Professor Kim Dae-soo of KAIST’s Department of Brain and Cognitive Sciences and Director Lee Chang-joon of the Institute for Basic Science (IBS) also took part in the project.
Parkinson's disease is a brain disorder that gets worse over time. It causes symptoms such as shaking, stiff muscles, slow movement, and trouble with balance. Famous figures like Mohammad Ali and Michael J. Fox have lived with it for years. Doctors often cannot detect it early because the first changes in the brain are too small for regular tests to catch. Current drugs that target brain signals also have limited success.
The KAIST-led team tried something different. They studied mice engineered to develop Parkinson's-like symptoms. Using several cameras and artificial intelligence, they tracked more than 340 types of body movement, such as walking patterns, hand and foot motions, and tremors. The AI system turned this complex information into one score, called the "Parkinson’s behavior index."
This score revealed early signs of Parkinson’s disease just two weeks after symptoms began. It was more accurate than older motor function tests. The most telling clues were uneven movement between left and right limbs, shorter walking steps, posture changes, and chest tremors.
To prove that the index was specific to Parkinson’s and not just general motor problems, the team tested it on mice with ALS, another disease that affects movement. Those mice did not show the same score, confirming that the new system really points to Parkinson's-specific changes.
For treatment, the researchers turned to a technique called optogenetics, which uses light to control the activity of brain cells. Their method, called optoRET, shone light on certain brain cells linked to dopamine, the chemical heavily affected by Parkinson’s disease. In the mouse experiments, this treatment improved walking, made limb movements smoother, and reduced tremors. The best results came when the light was used every other day.
"This study is the first in the world to bring AI-based behavior analysis and optogenetics together in one framework for Parkinson’s," said Professor Heo. "It provides a foundation for new therapies that can be customized for each patient."
The findings were published on August 21 in Nature Communications. KAIST researcher Hyun Bo-bae was the first author. She is now continuing related research at McLean Hospital, part of Harvard Medical School, with support from South Korea’s Korea Health Industry Development Institute.
The project was supported by KAIST’s Global Singularity Research Program, the Ministry of Science and ICT, the National Research Foundation of Korea, IBS, and the Ministry of Health and Welfare.
reporter
Park Sae-jin
swatchsjp@ajunews.com
Copyright ⓒ Aju Press All rights reserved.
