Case Study: Duke University accelerates Parkinson’s disease neuron quantification with Aiforia AI

Duke University achieves 99%+ F1-score for Parkinson’s research with Aiforia

Duke University, specifically a pharmacology and cancer biology lab, faced significant challenges in its Parkinson's disease research due to traditional, time-consuming manual methods for counting dopaminergic neurons. These methods were not only slow, taking weeks or months, but also prone to human error and bias. To overcome this, the lab turned to Aiforia and its Aiforia Create platform to build an AI model for automated cell quantification.

Using Aiforia, the team developed an AI model to quantify tyrosine hydroxylase (TH)-positive neurons. The solution proved to be incredibly fast, reducing analysis time from weeks or months to just a few hours, and extremely accurate, achieving an F1-score of over 99% when validated against human counting. This efficiency gain allowed the lab to expand its research scope. As a result of the success with Aiforia, the lab plans to incorporate this AI model into all future projects and is building additional models for other key markers.

Duke University