My PD Story

2025 Trailblazer Award 

Combining AI and Skin Biopsies to Detect Parkinson’s Years Earlier 

Parkinson’s disease (PD) is primarily diagnosed clinically, meaning that a diagnosis is based on an assessment of symptoms and medical history. While certain tests looking at PD biomarkers can help support a diagnosis, currently there is no specific lab test that can be used to diagnose PD by itself. Additionally, there is currently no way to efficiently predict who may be at risk for developing PD, which is critical for potential therapeutic interventions.  

Jay Alberts, PhD, recipient of a Parkinson’s Foundation Trailblazer Award, is leveraging an explainable artificial intelligence (AI) model to identify those at-risk for developing PD within two years based on medical history analysis. He will then invite those deemed “high-risk” to test out a new diagnostic test that could detect PD with a simple skin biopsy, improving our ability to detect the disease earlier. 

“This project leverages the power of artificial intelligence to provide greater understanding and use of a promising biomarker test in individuals who may be in the pre-symptomatic phase of PD.” – Dr. Alberts 

Alpha-synuclein, a protein whose biochemically altered form is prone to neuron-damaging clumping in PD, is widely agreed to be a strong candidate biomarker for the disease. Recent research has found that biochemically altered alpha-synuclein can readily be detected in skin samples of those with symptomatic PD. Dr. Alberts is working to see if the same is true of those with pre-symptomatic PD and if AI can best identify people who should undergo this testing. 

Dr. Alberts and his team at the Cleveland Clinic, a Parkinson’s Foundation Center of Excellence, have developed an AI program that can analyze health records to determine a person’s risk of developing PD within a few years. Importantly, the AI system is “explainable,” meaning that the model can inform patients on the relative contribution of each predicting factor that leads into their overall risk score.  

For his study, Dr. Alberts will invite people the AI model has deemed to be high and low risk of future PD to undergo a Syn-One test. This diagnostic test involves three small skin biopsies — each approximately 1/8 of a pencil eraser in size — from which levels of biochemically altered alpha-synuclein are measured. If Dr. Alberts’ hypothesis is correct, then those deemed high-risk for future PD will have higher amounts of this biomarker than those deemed low-risk.  

“This project will, for the first time, combine explainable artificial intelligence models developed from the electronic health record to identify and evaluate a scalable approach to screening for Parkinson's disease,” said Dr. Alberts.  

If these validation studies go well, this combination of AI-powered risk determination with Syn-One testing could be used to detect PD years earlier than currently possible. 

Meet more Parkinson’s researchers! Explore our My PD Stories featuring PD researchers.