Now, IBM is using machine learning to create a blood test that could possibly detect the disease decades before the symptoms even start. Doctors use a combination of medical history review, physical exam, neurological exam, "mental status" test and brain imaging to make the diagnosis.
IBM also claims that amid the wide range of other proposed blood tests for Alzheimer's disease that are now being developed, this is the first study to use machine learning to identify sets of proteins in the blood that are predictive of a biomarker in spinal fluid.
An early study found that tiny blood vessels that normally form a dense web at the back of the eye become sparser in Alzheimer's patients. Not only were the researchers able to detect differences between the Alzheimer's patients and the other two groups, but they were also able to see differences among the Alzheimer's patients that appeared to be linked with the severity of the disease.
The objective of the research is to determine whether the changes that occur in the brain of a person with Alzheimer's could be detected in the retina even before symptoms, and memory loss can occur.
The test involves looking at blood vessels in the retina, but isn't something that now takes place as part of a normal eye test.
For their study, researchers used OCTA to compare the retinas of 39 Alzheimer's patients with 37 people with mild cognitive impairment, and 133 cognitively healthy people.
Trump Makes Conspiracy Claim About 'Fake Melania' Conspiracy Theory
Popular US talk show The View also weighed in on the Fake Melania talk, and were later criticised by Stephanie Grisham, a spokesperson for Melania.
The differences seen were statistically significant after adjusting for factors such as age, sex and education. Within less than a minute, OCTA takes a non-invasive picture of the retina - the nerve cells in the back of our eyes that convert light from the outside world into nerve signals that the brain interprets into images.
"Cognitively normal, healthy individuals do not have these changes in their retina", explained Dr. Sharon Fekrat, an ophthalmologist at Duke Eye Center and author of the study.
In terms of how the test works, IBM's Ben Goudey says that the team uses machine learning in order to identify a specific biological marker for Alzheimer's called amyloid-beta. Their study was published online today in Ophthalmology Retina, a journal of the American Academy of Ophthalmology.
"The retina has a dense spider web of blood vessels, and we were interested in the density of these blood vessels". Prof Fekrat and colleagues said this offers "a window into the disease process".
"It's not possible for current techniques like a brain scan or lumbar puncture to screen the number of patients with this disease".