Machine learning reveals dementia type from brain scans

TBH LP - Image 4

By Alice Tuohy | Monday 22 October 2018

Nature Communications: Uncovering the heterogeneity and temporal complexity of neurodegenerative diseases with Subtype and Stage Inference (SuStaIn)

Researchers in the UK have used a powerful data-driven machine learning tool, called SuStaIn, to identify the exact form and stage of Alzheimer’s disease and frontotemporal dementia (FTD) from brain scans alone. The findings are published in the scientific journal, Nature Communications.

Alzheimer’s disease and FTD both cause dementia yet each have distinct genetic underpinnings and different underlying biological mechanisms. While both have unique symptoms, as the diseases get worse over time, symptoms can often overlap and make diagnosis even more difficult.

The research, led by scientists at University College London and part-funded by Alzheimer’s Research UK, used a machine learning approach to measure small changes in brain size in different regions of the brain.

Using an MRI image databases of 365 volunteers from 13 centres in the UK, SuStaIn scanned detailed pictures for small changes linked to disease and progression. The results reveal that the technique can recognise different types and stages of the two diseases and could aid a more accurate diagnosis of each disease.

Dr Sara Imarisio, Head of Research at Alzheimer’s Research UK, said:

“Diagnosing dementia accurately can be a challenge, particularly as many of the early symptoms may overlap with other health conditions. Machine learning is an incredibly powerful tool and we are only just beginning to realise its full potential to analyse vast and intricate datasets in dementia research.

“In this new study the value of machine learning is demonstrated through its ability to bring together brain scans and sophisticated computational algorithms to give in-depth insights into the subtle brain changes in Alzheimer’s disease and frontotemporal dementia.

“Before any new diagnostic techniques are used in the clinic we need to be confident that they won’t wrongly flag up healthy individuals or miss anyone who might have benefitted from intervention. The next steps for this research are to follow-up the findings with larger groups of people to understand more about how sensitive the technique is compared to other diagnostic approaches.

“While we need to see further studies building on these results, Alzheimer’s Research UK is pleased to have supported this pioneering research to take us closer to a more accurate diagnosis of the diseases that cause dementia.”


By submitting a comment you agree to our comments policy.

Leave a Comment

About the author

Alice Tuohy