Schizophrenia – a very serious disease, characterized by violation of coherence of mental processes, and decline of mental activity. The overall risk of the disease, according to the studies, ranging from 0.4 to 0.6%, that is approximately 4-6 cases per 1000 people. In America alone suffer from schizophrenia 3.2 million people, so the American scientists are trying to find a way how can early detection of the disease. Thanks to the efforts of experts at IBM and researchers from the University of Alberta, this method was found.
For early diagnosis of the symptoms of schizophrenia in humans doctors only need MRI scans of his brain. Pictures are uploaded in a specially designed IBM system, where the neural network is carefully studying the blood flow inside the brain. A team of researchers have trained an artificial intelligence with the help of MRI images 95 anonymous patients, some of whom were healthy and some are ill with schizophrenia. The photographs depict the blood flow in the time patients simple audioprojektai. Based on these data, the neural network has formed a predictive model able to detect the signs of schizophrenia in the studied patient. While testing the system on a specially selected test group, the neural network was able to identify schizophrenia patients in 74% of cases, which is an excellent result.
“This unique and in many respects innovative approach to the problem opens up new medical horizons. We are incredibly close to understanding the neurobiology of schizophrenia, which will help us in the future to improve treatment methods and control this terrible disease. We found inside the human brain, the plurality of markers that have been linked with schizophrenia, and we have created a neural network based on these patterns, will be an excellent diagnostic tool,” — shared his joy Professor of psychiatry and neuroscience Serdar Dursun.
Artificial intelligence has learned how to identify early symptoms of schizophrenia
Sergey Grey