A Groundbreaking Platform: AI Meets Optogenetics
KAIST researchers have unveiled a pioneering preclinical platform that fuses artificial intelligence (AI) with optogenetics to advance both the diagnosis and treatment of Parkinson’s disease. The announcement, made by KAIST president Kwang Hyung Lee on September 22, highlights a collaboration spanning the university’s Biological Sciences, Brain and Cognitive Sciences, and IBS Center for Cognition and Sociality, led by Professor Won Do Heo, Professor Daesoo Kim, and Director Chang-Jun Lee.
Analyzing 340 Behavioral Traits to Build APS
In a mouse model of Parkinson’s, the team analyzed more than 340 behavioral features, including gait, limb movements, and tremors, to derive a unique metric dubbed the APS score. The APS distinguished Parkinson’s models from controls as early as two weeks after disease onset, demonstrating greater sensitivity than traditional motor tests. The algorithm highlighted changes in gait, asymmetric limb movements, and chest tremors as key diagnostic signs.
Ensuring Specificity: Parkinson’s versus ALS
To confirm that these signatures were specific to Parkinson’s, the same analysis was applied to a mouse model of amyotrophic lateral sclerosis (ALS). Despite motor debility, ALS mice did not show elevated APS scores, underscoring the method’s disease specificity and reducing concerns that general motor decline could drive the results.
OptoRET: Light-Based Therapeutic Modulation
On the therapeutic side, researchers employed optoRET, an optogenetic technique that enables light-controlled delivery of neurotrophic signals in the brain. In the Parkinson’s mouse model, optoRET reduced tremors and lent a smoother quality to both gait and limb movements. A regimen of alternating-day light exposure proved most promising and even suggested a neuroprotective effect on dopaminergic neurons.
Implications for Personalized Medicine
According to Professor Won Do Heo, this work marks the first time a preclinical platform connects early diagnosis, therapeutic assessment, and mechanistic validation of Parkinson’s disease through integrated behavioral analysis with AI and optogenetics. The approach lays a crucial foundation for the future of personalized medicine, where diagnostics and targeted therapies could be tailored to individual neural and behavioral signatures.
Publication and People Behind the Work
The study, with Dr. Bobae Hyeon as first author, was published online in Nature Communications on August 21. Dr. Hyeon is a postdoctoral researcher currently pursuing additional work at McLean Hospital, Harvard Medical School, supported by the Global Physician-Scientist Training Program of the Korea Health Industry Development Institute.
Funding and Collaboration
Funding for the research came from the KAIST Global Singularity Project, the Ministry of Science and ICT/National Research Foundation of Korea, IBS Center for Cognition and Sociality, and the Ministry of Welfare and Health/Korea Health Industry Development Institute.
