Artificial intelligence can be used to detect Alzheimer's disease six years before a patient's overall diagnosis, according to a study.
Doctors used a self-paced computer to detect brain-scan changes that people seem to be too sophisticated.
The system was able to identify dementia in 40 patients, on average six years before the official diagnosis.
British AI expert Prof Nobela Sharkey of the University of Sheffield said of the findings: "It is precisely such a task that deep training is stopped – finding high-level data models.
"Although the sample sizes and test kits were relatively small, the results are so promising that a much larger study would be desirable."
Boffins from the University of California trained a computer using over 2100 scans from 1002 patients.
Scans the measurement of the brain's activity by tracking the intake of radioactive fluid in the blood.
Studies have linked Alzheimer's disease to specific changes in some of the brain regions, but these may be difficult to spot.
The Alzheimer's algorithm was able to teach you to recognize the brain scan patterns that point to the disease.
As a final test, it was given 40 scans from 40 patients that it had never previously studied.
Many years before the patient was diagnosed with Alzheimer's disease, it turned out to be 100 percent precise.
Dr Jae Ho Sohn, who worked on the project, said: "We are very pleased about the algorithm's performance.
"It was able to predict each case that has Alzheimer's disease."
Early detection of Alzheimer's disease could open up new opportunities for slowing or even stopping the progression of the disease.
Dr Carol Routledge of Alzheimer's Research UK said: "Diseases that cause dementia begin in the brain up to 20 years before symptoms begin to appear, providing a vital opportunity for us to intervene before widespread lesions occur.
"This study highlights the potential of machine learning to help detect diseases like Alzheimer's faster, but conclusions will need to be approved by much larger groups of people before we can properly evaluate the possibilities of this approach."
The research is published in the latest edition of the journal Radiology.
Click to get more from The Sun.