Artificial intelligence could help fast-track dementia diagnosis

ADVERTISEMENT

Various forms of dementia can be detected faster and easier by analyzing recordings of patients’ electrical brain activity using artificial intelligence (AI), according to new research.

Scientists from the University of Surrey and Newcastle University have shown that electroencephalography (EEG) can be used as a low-cost diagnostic tool to help doctors identify different forms of dementia, including Lewy body, Alzheimer’s disease and Parkinson’s dementia.

Our study shows that using AI analysis of EEG data as a diagnostic tool for identifying dementia can change the lives of many people. We have shown that by combining brain activity from patients with their eyes open and their eyes closed, our machine learning algorithms can accurately detect various forms of dementia, including Lewy body dementia, which is often only found after death. As a result, we believe our method could allow people to be diagnosed and treated sooner.

An obvious next step for our project is to gather support for a clinical trial of this incredibly promising technology.”

Dr Roman Bauer, first author of the study at the University of Surrey

In the study, researchers used EEG data (with eyes open and eyes closed) from 40 people with dementia in their 70s. In addition, the study used 15 health control subjects.

According to the World Health Organization, there are 55 million people living with dementia worldwide. However, studies have shown that Lewy body dementia can be found in more than 25% of cases of postmortem dementia, indicating an underrepresentation of this specific type in the current data.

The study has been published in Alzheimer’s Research & Therapy.

source:

Journal reference:

Jennings, J.L., et al. (2022) Investigation of the power of open eyes in a resting state EEG to aid in the diagnosis of dementia. Alzheimer’s disease research and treatment. doi.org/10.1186/s13195-022-01046-z.

ADVERTISEMENT

Leave a Reply

Your email address will not be published.