Researchers at the University of São Paulo in Brazil are using artificial intelligence (AI) and Twitter to try to create models that can predict anxiety and depression before clinical diagnosis.
Their goal is to detect the signs of these disorders before they become more severe.
The study started with the construction of a database called SetembroBR, named after Yellow September, an annual suicide awareness and prevention campaign.
The database contains information relating to a corpus of texts in Portuguese and the network of connections involving 3,900 Twitter users who reported having been diagnosed with or treated for mental health problems before the survey.
The corpus includes all public tweets posted by these users individually, for a total of around 47 million tweets.
The researchers collected timelines manually, analyzing tweets by some 19,000 users, equivalent to the population of a village or small town.
They then used two datasets, one for users who reported being diagnosed with a mental health problem and another selected at random for control purposes.
Their aim was to distinguish between people with depression and the general population.
The study also collected tweets from friends and followers of the users, as people with mental health problems tend to follow certain accounts, such as discussion forums, influencers, and celebrities who publicly acknowledge their depression.
Mental health disturbances, including depression and anxiety, are a growing global concern.
According to the World Health Organization, 3.8% of the world’s population, or some 280 million people, were affected by depression in 2021. WHO also estimated an increase of 25% in global prevalence of these mental health problems during the COVID-19 pandemic.
The tweets were collected for the study during this period.
To process the data, the researchers used deep learning, an AI technique that teaches computers to process data in a way inspired by the human brain.
They created four text classifiers and word embeddings using models based on bidirectional encoder representations from transformers (BERT), a machine learning algorithm for natural language processing.
These models correspond to a neural network that learns contexts and meanings by monitoring sequential data relationships, such as words in a sentence.
The training input consisted of a sample of 200 tweets selected at random from each user. The researchers found that BERT performed best in terms of predicting depression and anxiety.
People with depression tended to write about subjects connected to themselves, using verbs and phrases in the first person, as well as topics such as death, crisis, and psychology.
The researchers anonymized all the collected texts and didn’t publish actual tweets or users’ names to protect people’s identities.
They are now extending the database, refining their computational techniques, and upgrading the models in order to produce a tool for future use in screening prospective sufferers of mental health problems and helping families and friends of young people at risk from depression and anxiety.
Brazil ranks third among the countries that consume social media in the world, according to a Comscore survey published in early March.
Its 131.5 million users are online for 46 hours a month on average. The most widely used platforms are YouTube, Facebook, Instagram, TikTok, Kwai, and Twitter.
If you care about depression, please read studies that vegetarian diet may increase your depression risk, and Vitamin D could help reduce depression symptoms.
For more information about health, please see recent studies that ultra-processed foods may make you feel depressed, and these antioxidants could help reduce the risk of dementia.
The study was conducted by Wesley Ramos dos Santos et al and published in Language Resources and Evaluation.
Copyright © 2023 Knowridge Science Report. All rights reserved.