King’s College London King’s MND Care and Research Centre King’s College Hospital

Artificial intelligence for motor neurone disease

We build and evaluate AI systems for MND — helping families understand inherited risk, making clinical trials need fewer people on placebo, and shortening the path from genetic signal to discovery.

Our research Try the demo

Background

What motor neurone disease is

Motor neurone disease is a group of progressive neurodegenerative conditions that damage the motor neurones carrying signals from the brain and spinal cord to the muscles. As those neurones fail, people gradually lose the ability to move, speak, swallow and breathe. The most common form is amyotrophic lateral sclerosis, or ALS.

There is no cure. Care focuses on slowing progression where possible, managing symptoms, and supporting people and their families to live as well as they can for as long as they can.

1 in 300 Lifetime risk of developing MND

Up to 5,000 Adults living with MND in the UK at any one time

6 a day People diagnosed with MND in the UK

2–5 years Typical survival from first symptoms in ALS, the most common form

Figures from the MND Association.

Why genetic counselling matters here

Most cases of MND are not inherited. But some families do carry a genetic cause, and as testing becomes more widely available, more relatives face difficult questions: whether to be tested at all, what a result would and would not mean, and how to talk about it with the rest of the family.

Genetic counselling exists to help people work through exactly those questions, at their own pace and without being rushed. Specialist counselling capacity is limited, and demand is growing. That gap is where our first project began.

Programme

Our research

Three strands of work, with patient and public involvement running through all of them.

Project 01

An LLM agent for genetic counselling

A conversational agent that helps people affected by MND understand what inherited risk does and does not mean, and that supports genetic counsellors around a consultation rather than replacing them. A working demo exists today — we would welcome people trying it.

Project 02

Digital twins instead of placebo arms

AI models that predict how a given participant would have progressed without treatment, creating a digital twin that can stand in for a placebo control. If the prediction holds up, fewer people need to be randomised to placebo — and more of them can receive the candidate drug.

Project 03

LLM agents to accelerate genetics discovery

Agents that read, connect and reason across genomic data and the published literature, shortening the path from a genetic signal to a hypothesis worth testing at the bench or in the clinic.

Throughout

Patient and public involvement

People living with MND, and the families around them, help set our priorities, shape how studies are designed, and judge whether what we build is worth using. PPI runs alongside every project above rather than being consulted once at the end.

Principles

How we work

In order of what comes first.

01 Co-design with clinicians and patientsClinicians and people affected by MND are in the room from the first design meeting, not shown a finished product and asked what they think of it.
02 Strict evaluation, from in silico to real-world clinical studiesEvery system is tested in simulation, then against retrospective data, then in prospective clinical studies. We report where it fails, not only where it works.
03 Transparent, local, open-weight modelsOpen-weight large language models we run ourselves are our core method. Nothing we cannot inspect, and no dependence on an external vendor’s API for clinical work.
04 Governed dataNo data leaves King’s or the hospital. The models come to the data, under King’s College London and NHS information governance.

People

The team

Taiyu Zhu Lecturer in Large Language Models for Healthcare Department of Biostatistics & Health Informatics

Ammar Al-Chalabi Professor of Neurology and Complex Disease Genetics; Director of the King’s MND Care and Research Centre Department of Basic & Clinical Neuroscience

Alfredo Iacoangeli Professor of Bioinformatics and Artificial Intelligence for Genomic Medicine Departments of Basic & Clinical Neuroscience and Biostatistics & Health Informatics

Get in touch

Contact

Try the genetic counselling demo

We have a working demo of the genetic counselling agent. If you are a clinician, a researcher, or someone affected by MND and you would like to try it, please email us — we would be glad to hear from you and to hear what you make of it.

taiyu.zhu@kcl.ac.uk

Research enquiries

taiyu.zhu@kcl.ac.uk

Collaborations, data requests, students and press.

Where we are

Department of Biostatistics & Health Informatics

School of Mental Health & Psychological Sciences
King’s College London

Looking for support with MND rather than research? The MND Association provides information, local support and a free helpline. Your GP or neurologist can refer you to a specialist MND care centre.

Supported by

This research is supported by the organisations above. The views expressed on this site are those of the researchers and not necessarily those of the funders.