research
<Research team photo (from top left) Dr. Bobae Hyeon, Professor Daesoo Kim, Director Chang-joon Lee, (right) Professor Won Do Heo>
Globally recognized figures like Muhammad Ali and Michael J. Fox have long suffered from Parkinson's disease. The disease presents a complex set of motor symptoms, including tremors, rigidity, bradykinesia, and postural instability. However, traditional diagnostic methods have struggled to sensitively detect changes in the early stages, and drugs targeting brain signal regulation have had limited clinical effectiveness.
Recently, Korean researchers successfully demonstrated the potential of a technology that integrates AI and optogenetics as a tool for precise diagnosis and therapeutic evaluation of Parkinson's disease in mice. They have also proposed a strategy for developing next-generation personalized treatments.
KAIST (President Kwang Hyung Lee) announced on the 22nd of September that a collaborative research team—comprising Professor Won Do Heo's team from the Department of Biological Sciences, Professor Daesoo Kim's team from the Department of Brain and Cognitive Sciences, and Director Chang-Jun Lee's team from the Institute for Basic Science (IBS) Center for Cognition and Sociality—achieved a preclinical research breakthrough by combining AI analysis with optogenetics. Their work simultaneously demonstrated the possibility of early and precise diagnosis and treatment in an animal model of Parkinson's disease.
The research team created a Parkinson's disease mouse model with two stages of severity. These were male mice with alpha-synuclein protein abnormalities, a standard model used to simulate human Parkinson's disease for diagnostic and therapeutic research.
In collaboration with Professor Kim's team at KAIST, they introduced AI-based 3D pose estimation for behavioral analysis. The team analyzed over 340 behavioral features—such as gait, limb movements, and tremors—from the Parkinson's mice and condensed them into a single metric: the AI-predicted Parkinson's disease score (APS).
The analysis showed that the APS exhibited a significant difference from the control group as early as two weeks after the disease was induced. It also proved more sensitive in assessing the disease's severity than traditional motor function tests. The study identified key diagnostic features, including changes in stride, asymmetrical limb movements, and chest tremors. The top 20 behavioral features included hand/foot asymmetry, changes in stride and posture, and an increase in high-frequency chest movement.
To confirm that these behavioral indicators were not just general motor decline but specific to Parkinson's, the team applied the same analysis to a mouse model of Amyotrophic Lateral Sclerosis (ALS), also known as Lou Gehrig's disease, in collaboration with Director Lee's team at IBS. Since both Parkinson's and ALS cause motor function problems, if the APS simply reflected poor motor skills, a high score should have appeared in both diseases.
However, the analysis of the ALS animal model showed that despite a decline in motor function, the mice did not exhibit the high APS seen in the Parkinson's model. Instead, their scores remained low, and their behavioral changes were distinctly different. This demonstrates that APS is directly related to specific, characteristic changes that only appear in Parkinson's disease.
For treatment, the research team used optoRET, an optogenetics technology that precisely controls neurotrophic signals with light. This technique proved effective in the animal model, leading to smoother gait and limb movements and a reduction in tremors.
Specifically, a regimen of shining light on alternate days was found to be the most effective, and it also showed a tendency to protect dopamine-producing neurons in the brain.
Professor Won Do Heo of KAIST stated, "This is the first time in the world that a preclinical framework has been implemented that connects early diagnosis, treatment evaluation, and mechanism verification of Parkinson's disease by combining AI-based behavioral analysis with optogenetics." He added, "This lays a crucial foundation for future personalized medicine and customized treatments for patients."
The study, with Dr. Bobae Hyeon, a postdoctoral researcher at the KAIST Institute for Biological Science, as the first author, was published online in the international journal Nature Communications on August 21st. Dr. Hyeon is conducting follow-up research to advance Parkinson's cell therapy at McLean Hospital, Harvard Medical School, supported by the "Global Physician-Scientist Training Program" of the Korea Health Industry Development Institute.
This research was supported by the KAIST Global Singularity Project, the Ministry of Science and ICT/National Research Foundation of Korea, the IBS Center for Cognition and Sociality, and the Ministry of Health and Welfare/Korea Health Industry Development Institute.
- Paper Title: Integrating artificial intelligence and optogenetics for Parkinson's disease diagnosis and therapeutics in male mice
- DOI: https://doi.org/10.1038/s41467-025-63025-w
-
research KAIST Automates the Search for “Dream Semiconductor” 2D Semiconductors
The era of researchers manually searching for two-dimensional semiconductors, which are drawing attention as next-generation AI semiconductors, is coming to an end. KAIST researchers have automated semiconductor screening and device fabrication, analyzed thousands of devices, and revealed the relationship between thickness and performance that had long been difficult to identify. This achievement is expected to shift next-generation semiconductor research toward a data-driven approach and acce
2026-07-09 -
research KAIST Identifies the “Hidden Energy Cost” of AI Agents for the First Time
As the era of AI agents—systems that can reason and act autonomously—begins, the power consumption of data centers is emerging as a critical challenge. A KAIST research team has, for the first time, analyzed the computational cost and energy consumption of AI agents, finding that they can consume up to 136.5 times energy per query than conventional generative AI. The study shows that competitiveness in the AI era is expanding beyond model performance to include the efficiency of d
2026-07-07 -
event KAIST Global Entrepreneurship Summer School Marks Fifth Consecutive Year of Cultivating Future Entrepreneurs in Silicon Valley
The 2026 Global Entrepreneurship Summer School (GESS), organized by the KAIST Office of Global Initiatives, has successfully concluded its fifth annual program. Now in its fifth year, GESS has become KAIST's flagship global entrepreneurship program, providing students with firsthand experience in Silicon Valley—the world's leading startup ecosystem—and equipping them with the entrepreneurial mindset and global competencies needed to launch ventures on the international stage.
2026-07-06 -
research KAIST Develops AI That Reads Animal Behavior Like Language
An artificial intelligence model capable of reading and interpreting animal behavior like language has been developed by researchers at KAIST. The team created BehaVERT, an AI model that learns behavioral data in a manner similar to natural language and was able to independently identify social behavioral deficits in an autism mouse model, opening a new avenue for interpretable neuroscience. KAIST (President Kwang-Hyung Lee) announced that a research team led by Professor Dae-Soo Kim from the
2026-07-01 -
research KAIST Identifies Hidden Age Bias in Artificial Intelligence
Do responses generated by artificial intelligence systems such as ChatGPT reflect social prejudice? A KAIST research team has quantitatively analyzed and identified age-related stereotypes embedded in the responses of generative artificial intelligence. The study sheds light on the potential impact of hidden AI biases on social perceptions and suggests directions for the development of more inclusive AI. KAIST, led by President Kwang Hyung Lee, announced on the 28th that a research team led b
2026-07-01