Study: Your Publications on Facebook Reveal the State of Your Mental Health
Researchers have developed an algorithm based on artificial intelligence, which they believe can scan people's publications on their accounts via social media, and warn them if there are symptoms of mental illness, according to Russia Today. Symptoms included signs of hostility, loneliness and words such as "tears" and "emotions" as well as excessive use of self and the pronoun "I".
"What people write about social media and the Internet embodies a very difficult aspect of life in medicine and research," said Andrew Schwartz, lead author and lead researcher at the University of Pennsylvania, Stanford University. Access ". "It is relatively untapped compared to the biophysical markers of the disease," Schwartz said. "For conditions like depression, anxiety and post-traumatic stress disorder, for example, there are more signs in the way people express themselves digitally."
The social media data contained signs closer to the genome, as the data had surprisingly similar methods to those used in genomics, so social data acquisition could be an effective way to find early signs of mental illness, the research team said. Depression seems to be one of those diseases that can be discovered in this way, as people change their way of using social media. A total of 1,200 participants in the study agreed to provide a digital archive for their use of Facebook and their medical records. Of those, only 114 were depressed in their medical records, and each was compared with 5 non-depressed individuals to test the accuracy of the program.
By analyzing 524,292 participants who were published by Facebook volunteers in their pre-depression years, the team identified "language signs associated with depression." The researchers identified the most frequently used words and phrases and then put up 200 topics to explain what they called "language signs associated with depression" that allowed the program to detect early signs of depression in individuals by publishing it on Facebook before being officially diagnosed in their medical records by three Months.
The study also found that its program was more accurate in diagnosing depression, using social media signals in the six months prior to the onset of signs of the disease. The research team explained that these early signs of the disease include emotional, cognitive and personal processes such as enmity, loneliness, sadness and meditation. This can help to report the onset of depression in people at risk, enabling the early diagnosis and treatment of the disease, which reduces its impact on education, employment and social relations. Researchers say the program can be improved by integrating phone use data or face recognition software to analyze images published on Facebook.