Researchers reveal that depression can be identified through the content and style of a person’s language, according to a new study published on February 6, 2018.
Scientists have unveiled a class of words that accurately predict whether someone is suffering from depression. They made use of computerized text analysis methods to process extremely large data of words in minutes. This system helped spot linguistic features that could be missed out by humans, further calculating the prevalence of words and classes of words, lexical diversity, average sentence length, grammatical patterns, and other metrics.
Researchers have studied personal essays and diary entries of depressed people of well-known artists such as Cobain and Plath. They found consistency in language of people with symptoms of depression, striking a clear difference between usage of words of those with and without depression.
Two important factors such as content and style of language were studied. Expression in words is what was studied as content, this revealed the meaning of the statement expressed. They found that people suffering from depression—reveled the most expected results—excessively use words conveying negative emotions such as ‘lonely’, ‘sad’ or ‘miserable’. They further found that they were more inclined to using singular personal pronouns such as ‘me’, ‘myself’ and ‘I’ and significantly fewer second and third person pronouns. This suggests that people with depression are more focused on themselves and less connected with others.
According to a big data text analysis of 64 different online mental health forums, which examined over 6,400 members, reveals that absolutist words were more commonly used by mental health forums to identify depression. Absolute words convey absolute magnitudes or probabilities such as ‘always’, ‘nothing’ or ‘completely’ mental health forums than either pronouns or negative emotion words.
Examining 19 different control forums shows that the prevalence of absolutist words is around 50% greater in anxiety and depression forums, and around 80% greater for suicidal ideation forums.