Scientists have learned to recognize depression using artificial intelligence
[ad_1]
A computer neural network was trained to distinguish really sick people from healthy ones using MRI images
Diagnose depressive disorder with an accuracy of 82.6% learned scientists from the Baltic Federal University. I. Kant on certain indicators of the nervous system of patients.
Major depressive disorder, which is associated with loss of interest in activities, insomnia, drowsiness and guilt, affects about 280 million people worldwide.
However, it is often difficult to make an accurate diagnosis. After all, the doctor most often relies on a subjective assessment of his condition by the patient. They decided to call on machine learning algorithms, that is, artificial intelligence, to help, but it had to be first trained on a large number of truly sick and healthy people.
To do this, the authors used functional magnetic resonance imaging to take pictures of the brain activity of already identified patients with depression and healthy people, and then built functional networks based on them. After the picture, reflecting the interaction of different parts of the brain, the intensity of blood flow in them, a classification of sick and healthy people was built.
On the basis of the obtained networks, the scientists compared the brain function in 35 patients with depressive disorder and in 50 healthy people, and then, using machine learning methods, they separated the signs of both groups.
The proposed approach made it possible to distinguish sick and healthy people with an accuracy of 82.6%.
[ad_2]
Source link
