Bipolar disorder, a disease with alternating episodes of euphoria and moments of great depression, can be discovered early by analyzing the posts on Twitter
A study of the National Tsing Hua University of Taiwan published on the website arXiv, according to which with the help of artificial intelligence we can see the signs of the disease years before the diagnosis.
The beginning of bipolar disorder, the authors emphasize, is characterized by symptoms such as excessive talkativeness, problems sleeping and rapid mood changes, and many patients share on social their condition, including the diagnosis.
To verify if it was possible to catch the warning signs the researchers analyzed 10 million tweets from 2006 to 2016 of more than 400 people who had had a diagnosis of bipolar disorder, and compared them with 400 other people taken at random and used as a control group.
The team analyzed the content, the frequency and the time of the post, also studying the variation in characteristics in time, ‘teaching’ and finally to an algorithm to perform the analysis and to distinguish between persons with the symptoms and healthy individuals. The method, the authors conclude, can identify 90% of patients with up to one year prior to the diagnosis.
“The tweets of a person who suffers from bipolar disorder can provide a lot of information on the state of mind,” concludes the article.
These people are unconsciously providing a set of data that can be used to obtain information”. .