Shulou(Shulou.com)11/24 Report--

On the morning of August 16, Beijing time, it is reported that scientists have recently developed an AI algorithm that can detect early symptoms of patients with epilepsy, even symptoms that may be ignored by experienced doctors.

The results showed that the AI algorithm found 538 FCD symptoms during the scan, including 112 cases that could not be detected by radiologists.

The AI algorithm can detect patients with a rare disease with an accuracy of more than 60 per cent, compared with no hidden symptoms detected by MRI. The rare disease is epilepsy, which is said to have an incidence of 1% in British and American residents, which can cause uncontrolled brain currents in the brain, causing spasms and convulsions.

Anyone can have convulsive seizures, but that doesn't mean they all have epilepsy. Usually people with epilepsy may have convulsions more than once before they are diagnosed. When the brain sudden brain current burst, there will be epilepsy, resulting in brain operation interruption, some patients are still alert during seizures, can perceive the surrounding environment, while some patients will completely lose consciousness during seizures, in an extremely dangerous situation.

Sometimes patients with epilepsy have unusual sensations, cognition and activities, or their bodies are stiff and fall on the floor and keep twitching. At any age, stroke, brain infection, head injury or hypoxia at birth may induce epilepsy. In more than 50% of cases, doctors cannot find the specific cause, and related epileptic drugs cannot be completely cured. However, it helps to stop or reduce seizures, and if drug treatment is still ineffective, patients can only choose brain surgery.

One cause of epilepsy is drug-resistant focal cortical dysplasia (FCD), a subtle abnormality in the brain that causes brain signal transmission to fail. Epilepsy can be treated by surgery, but the changes in the brain caused by the disease are so subtle that even experienced radiologists may ignore the signs of the disease when performing MRI scans.

However, based on an AI algorithm developed by the research team of University College London, 63% of FCD symptoms can be detected, which is difficult for health care workers to detect before, and is also one of the important factors that induce seizures.

The researchers say their AI model can provide the best chance of cure for brain surgery for more patients with epilepsy. It is reported that in the UK, there are about 600,000 people with epilepsy, but only 20-30% of patients do not respond to drugs.

Brain cells or neurons usually form an organized layer of cells, forming the cerebral cortex. In patients with FCD, the brain cells are disordered, which leads to a higher risk of spasms and convulsions. FCD symptoms are the most common cause of epilepsy in children who receive surgical control, and FCD is the third most common cause in adults who need surgery. However, it is difficult for medical staff to detect FCD symptoms through MRI scans. in the latest study, researchers collected more than 1000 MRI scans from 22 global epilepsy studies. A team of radiologists marked the scans as healthy or FCD symptoms, and then ran the AI algorithm to detect scan abnormalities.

The AI algorithm, which involved information about 300000 regions of the patient's brain, published in the journal brain, showed that the AI algorithm detected 538 FCD symptoms during the scan, including 112 cases that radiologists could not detect.

The researchers say this is very important! Because the current symptoms of focal cortical dysplasia (FCD) mainly rely on timely detection and detection, and then choose surgical treatment. Dr Conrad Wagsteel (Konrad Wagstyl), a researcher at the Queen's Square Neurology Institute at University College London, said: "this AI algorithm helps to find more hidden symptoms of epilepsy in children and adults, prompting more epileptic patients to consider brain surgery as soon as possible to cure epilepsy and improve people's awareness of the disease. According to statistics, about 440 children undergo epileptic surgery each year. The condition improved after the operation. "

The AI algorithm is suitable for any patients with potential FCD symptoms, and the patients are over 3 years old and have undergone MRI scans. Co-author Dr Hannah Hannah Spitzer, a machine learning researcher at the Helmholtz Institute in Munich, Germany, said: "our algorithm has the ability to automatically learn to detect lesions from MRI data of thousands of patients. it can reliably detect lesions of different types, shapes and sizes, even many lesions that have previously been ignored by radiologists."

Dr Sophie Adler (Sophie Adler), Institute of Child Health at Greater Ormond Street, University College London, said: "We hope this technique will help detect abnormal symptoms that are easily overlooked and lead to epilepsy, which will eventually enable more people with epilepsy to recover through brain surgery."

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.