[ad_1]
The sooner you discover Alzheimer's disease, the better the treatment or intervention will be. Yet, it is difficult to detect the disease at an early stage. The diagnosis is often made only at an advanced stage because all the symptoms are present. The brain can then already be severely reduced, which sometimes makes the treatment unattractive.
Before the onset of the first symptoms, the uptake of glucose changes in brain cells. Although those who have "the naked eye" are hard to see, American researchers have now come up with a way to track these changes with the help of artificial intelligence.
Scientists from the University of California The Department of Radiology and Biomedical Imaging in San Francisco (UCSF) used in-depth learning techniques [19659004]. The computer then learns to recognize patterns based on a large number of examples.
These examples consisted of more than 2,000 FDG-PET scans of about 900 patients with Alzheimer's disease or mild cognitive impairment (MCI). With MCI, you also experience memory problems. The scans were collected over twelve years. Before participants entered the scanner, a type of radioactive glucose was injected into the blood. On the brain scanner, it is easy to see if glucose is absorbed by brain cells. The researchers examined the entire brain.
They then published the algorithm on forty subjects who had not yet developed Alzheimer's disease or another form of dementia. After six years on average, participants returned and a possible diagnosis was made. All participants who would receive Alzheimer's disease or ICM according to the algorithm have actually been diagnosed. 82% of subjects who remain healthy have not been diagnosed.
Subsequent research must test a larger group, write the researchers in Radiology. The algorithm could possibly detect in the future patterns related to the formation of beta-amyloid and tau tangles, features of Alzheimer's disease.
Source link
