Using AI to diagnose Parkinson's disease from the eye: the OCTAHEDRON project

Degenerative brain diseases such as Parkinson's and Alzheimer's can be most effectively treated when they are diagnosed early. Yet, the first changes occur deep inside the brain without any noticeable symptoms. To detect them early, we need to look at the retina, the light-sensitive tissue at the back of the eye - an outgrowth of the brain. This is now possible with a new imaging technique known as OCT (optical coherence tomography), which reveals microscopic details of the retina 3D structure.

In this talk, I will describe the OCTAHEDRON project that aims to use machine learning to detect early signs of neurodegenerative diseases in OCT scans. I will discuss how we use the NHS image data to train models to recognise different layers of the retina, and how these models can help in efficient diagnosis and early disease detection.