PhD student Misha Ramesh

Misha is a PhD student at the Alzheimer’s Society DTC for Vascular and Immune contributors to dementia. Her project is focused on:

Using Artificial Intelligence to analyse eye scans and look for signs of dementia

Thousands of people get images of their eyes (so called ‘retinal images’) taken each year, in community settings such as the opticians or at a hospital.

As retinal scans are relatively easy to obtain, I believe that if a good predictive model can be developed, this research has real potential to be translated into the healthcare setting and help with early diagnosis and treatment.

Tell us about your project 

The first part of my research project is about using an AI tool to analyse retinal images to get a better understanding of the human vascular system. Using these images is a non-invasive method of finding out the dimensions of blood vessels in the eye as a potential marker of blood vessels elsewhere in the body, particularly in the brain.

As the retinal vascular system and the brain vascular system are closely related, research has shown that several signs of neurodegeneration can be found in the retinal blood vessels. Neurodegenerative status can also be assessed using cognitive tests. The overall aim of my project is to build a statistical model which can predict the participants’ cognitive status based on their retinal blood vessel measurements, which could be an early sign of dementia.

Previous research has mainly focused on using these measurements to detect disease, such as Alzheimer’s, comparing those who were already diagnosed with those who did not have the disease. However, this research is aimed at seeing if we can predict the disease based on cognitive tests and retinal images that were taken before disease onset and diagnosis.

How will this research impact people living with dementia?

Retinal imaging is becoming increasingly available, particularly in ‘high-street’ opticians. Using AI tools to analyse retinal images allows detailed information about vessel size and shape to be extracted from large numbers of retinal images in a fully automated way.

These measures can be used to develop statistical models that can predict a person’s cognitive ability before a diagnosis is typically made. This would allow clinicians to provide more personalised interventions at an earlier stage than previously possible, which could in the future avert or delay the neurodegenerative process.

Overall, the research team and I aim to create a pipeline in which retinal images are automatically analysed using AI. The prediction model can then be used to predict cognitive status and allow clinicians to identify individuals who are more likely to have problems with their memory and thinking skills. These individuals could then be put forward for further testing to confirm a potential diagnosis and get access to support and treatments earlier.

What does it mean to you to be a part of an Alzheimer's Society Doctoral Training Centre?

Being a part of an Alzheimer’s Society Doctoral Training Centre is a rare opportunity for researchers early in their careers. It is a great privilege to have the opportunity to network with other PhD students and researchers who have an interest in dementia. It will allow me to learn new skills and gain a wider understanding of the field of dementia research.