Advertisement

Study: AI can help diagnose mental health disorders where access to care lacking

Oct. 15 (UPI) -- Artificial intelligence, or AI, may be able to screen people for mental health disorders without the need for a specialist's assessment, a study published Friday by GigaScience found.

In the analysis, the machine learning approach developed by researchers at the French Institute for Research in Computer Science and Automation, which relies on "proxy measures" for mental health, provided an accurate diagnosis up to about 90% of the time, the data showed.

Advertisement

Based on their findings, psychologists and machine learning models could work hand-in-hand in the future to provide personalized mental assessments, the researchers said.

For example, clients or patients could grant a machine learning model secured access to their social media accounts or their mobile phone data to then return useful proxy measures to aid specialists in diagnosis, they said.

However, while AI can provide much needed assessment tools, human interaction will still be essential, according to the researchers.

Advertisement

Other AI-based approaches have used a person's social media data to diagnose mental health problems.

"What is not going to change is that mental health practitioners will need to carefully interpret and contextualize test results on a case-by-case basis and through social interaction," lead researcher Denis Engemann said in a press release.

This would be the case "whether they are obtained using machine learning or classical testing," said Engemann, an experimental psychologist and research scientist at the French Institute for Research in Computer Science and Automation.

The prevalence of mental health and substance abuse disorders increased 13% globally between 2007 and 2017, the most recent year with figures available, according to the World Health Organization.

Many people around the world lack access to specialists who can provide a diagnosis and guide treatment, Engemann and his colleagues said.

Machine learning technology designed to facilitate mental-health assessments could provide much needed alternatives to help detect, prevent and treat such health issues, they said.

To develop AI models sensitive to mental health, Engemann and his colleagues used information for more than 500,000 adults from the U.K. Biobank, a database of medical reports and questionnaire responses gauging personal statistics and behaviors.

In addition to information on patient age, education, tobacco and alcohol use, sleep duration and physical exercise, questionnaire responses used in this study also included sociodemographic and behavioral data, such as moods and sentiments.

Advertisement

The data also included magnetic resonance imaging, or MRI, brain scans for more than 10,000 participants, the researchers said.

The researchers combined these two data sources to build models that approximate measures for brain age, and scientifically defined intelligence and neuroticism traits, or a person's tendency toward anxiety, depression, self-doubt and other negative feelings.

These characteristics served as proxy measures, which are indirect measurements that strongly correlate with specific diseases or outcomes that cannot be measured directly, according to the researchers.

They assessed the effectiveness of an AI algorithm in spotting mental health disorders using data for 7,000 U.K. Biobank participants, about half of whom had been diagnosed with a mental health disorder.

Using the combination of sociodemographic information and brain imaging to assess brain age -- a measure of cognitive function and performance -- provided the most accurate proxy measure for mental health, the data showed.

"We demonstrated that, beyond biological aging, the same proxy-measure framework is applicable to constructs more directly related to mental health," Engemann said.

"Second, we showed that useful proxy measures can be derived from other inputs than brain images, such as sociodemographic and behavioral data," he said.

The researchers validated their proxy measures by demonstrating the same results in a separate subset of UK Biobank data.

Advertisement

Latest Headlines