Research indicates that individuals with depression often express their negative emotions through writing and speaking. Surprisingly, a recent study published in the Proceedings of the National Academy of Sciences reveals that linguistic markers associated with depression are not prominent in the social media posts of Black individuals.
Efforts have been made by researchers and public health officials to utilize machine learning algorithms to detect potential connections between language patterns and mental health issues. These programs could serve as an early detection system by analyzing social media content to identify signs of depression within a specific population.
However, the latest research suggests that these AI programs may overlook signs of depression in a significant portion of the population. This discovery could have significant implications for public health, according to De Choudhury.
For their study, computer scientist Sunny Rai and her team at the University of Pennsylvania enlisted 868 participants in the United States, half of whom were Black and half were white. Participants completed a standardized depression survey online and granted access to their Facebook posts for analysis. The researchers then used a text analysis tool to examine the content of these social media posts.
Consistent with previous studies, the use of first-person singular pronouns like “I,” “me,” and “my” increased in correlation with depression scores across the entire group. Conversely, the use of first-person plural pronouns such as “we,” “our,” and “us” was associated with lower depression scores. Additionally, words reflecting negative emotions, such as those related to emptiness, disgust, and self-criticism, increased as depression scores rose.
2024-04-22 07:00:00
Read more on www.sciencenews.org