Large Language Model for Understanding and Monitoring Elderly Neurocognition

Dementia is a serious condition that affects millions of people, and getting a diagnosis early is really important for accessing the right care and support. Often, doctors learn about changes in memory and thinking from family members or carers – this is called a collateral history. However, these conversations can sometimes be tricky. Carers might find it hard to remember everything, explain symptoms clearly, or even just find the time to give a full picture. This can make the process of diagnosing dementia longer and more challenging.

This study is looking at a new AI (Artificial Intelligence) tool called LUMEN. Think of LUMEN as a helpful conversation guide. It's designed to assist carers in describing their loved one's symptoms more effectively. By gently guiding the conversation with clear questions, LUMEN aims to help carers put together a detailed and organised report for the doctors at memory clinics. The main goal here is to make the process of gathering information quicker, more accurate, and less stressful for everyone involved.

We will be testing LUMEN in real memory clinics across the UK. Carers and doctors will have the chance to use it and give us their feedback. We want to understand if it's easy to use, whether the information it gathers is better, and how it could fit into the way memory clinics currently work. We’re also holding workshops with different community groups to make sure the tool is suitable and accessible for people from all walks of life, regardless of their cultural background or the language they speak. If LUMEN proves to be successful, it could mean faster diagnoses, less pressure on carers, and ultimately, more people getting the dementia support they need sooner.