Scientists Develop AI Algorithms That Can Detect Brain Abnormalities, New Hope For Epilepsy Treatment

Add as a preferred source on Google

Researchers from all over the world, led by University College London (UCL), have developed artificial intelligence (AI) algorithms to detect brain abnormalities that cause epileptic seizures. [1]

The Multicentre Epilepsy Lesion Detection Project (MELD) algorithm was made by using MRI scans of more than 1,000 patients from 22 international epilepsy centres to find out where the problems are in drug-resistant focal cortical dysplasia (FCD), which is one of the main causes of epilepsy.

Take a look at the details.

About The Study

Epilepsy is a medical condition that causes the brain to work in strange ways, which can cause strange behaviour and seizures. [2]

FCDs are brain regions that have evolved incorrectly in drug-resistant epilepsy; It is said to be one of the common causes of this epilepsy type. Surgery is often used to treat it, but an MRI makes it hard for doctors to find the lesions.

In the study, the researchers measured cortical parameters using magnetic resonance imaging (MRI) in almost 300,000 locations throughout the brain to identify the thickness and fold of the brain area.

Based on their patterns and features, the algorithm was then trained on examples classified by qualified radiologists as what having an FCD brain or a healthy brain is.

What Was The Result?

According to the findings published in Brain, the algorithm correctly identified the FCD in 67% of the cases (538 participants) in the cohort.

Previously, radiologists were unable to detect the anomalies in 178 of the patients based on their MRI data. However, the MELD algorithm detected the FCD in 63% of these cases.

This is important because if the brain scan shows the abnormality, its treatment or surgery can be planned at an early stage or before the occurrence of any complications.

The researchers focused more on designing an AI system that was interpretable and could assist clinicians in making decisions. The algorithms of AI might make it easier to identify concealed lesions in epileptic children and adults. It is also designed to enable more epileptic patients to consider having brain surgery to cure their condition and speed up their cognitive development goals.

About Epilepsy

Epilepsy is a severe neurological disorder characterised by recurring seizures that affect 1% of the world's population. Around 600,000 people in the United Kingdom are affected. Although medications can treat the majority of epilepsy patients, 20-30% do not benefit from them.

"Our technology automatically learns to detect lesions from thousands of patient MRI scans," said Dr Hannah Spitzer, a co-first author from Helmholtz Munich. The AI can correctly identify lesions of varied types, shapes and sizes, including some that radiologists had previously overlooked.

"We believe that this technology may assist in discovering abnormalities that are now being missed that cause epilepsy," said Dr Sophie Adler, a co-senior author from the University College London's Great Ormond Street Institute of Child Health. According to her, in the long run, it may allow more epilepsy sufferers to undergo potentially curative brain surgery.

To Conclude

This FCD detection study employs the largest MRI cohort of FCDs to date, allowing it to identify all FCD subtypes.

The 22 hospitals that participated in the study used MRI scanners from around the world, which increased the algorithm's robustness but may have impacted its sensitivity and specificity.

Disclaimer: The information provided in this article is for general informational and educational purposes only and is not intended as a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or a qualified healthcare provider with any questions you may have regarding a medical condition.